INVESTIGATION OF IRON CONCENTRATION, VALENCE, AND SOURCE EFFECTS ON
SHORT-TERM IRON (Fe) STATUS AND DRINKING WATER PREFERENCE OF
LACTATING DAIRY COWS
By
Olivia N. Genther

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Animal Science
2011

ABSTRACT
INVESTIGATION OF IRON CONCENTRATION, VALENCE, AND SOURCE ON
SHORT-TERM IRON (Fe) STATUS AND DRINKING WATER PREFERENCE OF
LACTATING DAIRY COWS
By
Olivia N. Genther
2+

Drinking water can contain high concentrations of ferrous iron (Fe ), which is more
3+

absorbable than ferric iron (Fe ) found in feed and may negatively impact cow health and
productivity. Our objective was to characterize Fe status, oxidative stress, and drinking water
preference of lactating dairy cows in response to different concentrations, valences and sources
(salts) of Fe. In two dosing experiments, cows were abomasally administered 0, 0.75, or 1.5 mg
Fe/kg bodyweight from ferrous lactate; or, 0 or 1.5 mg Fe/kg bodyweight from ferrous or ferric
sulfate. Blood was sampled hourly for 12 h, and liver biopsies were taken at 0, 18 and 36 h postdosing. The Fe dosed had minimal effects on Fe status or oxidative stress, suggesting that
amounts of Fe administered do not affect short-term Fe metabolism. A protocol was developed
to evaluate water preference and drinking behavior of lactating cows for subsequent experiments.
In four preference experiments, cows were offered pairs of drinking water treatments. Water
intake was measured to determine preference for water with Fe of different concentrations,
valences and sources. Water intake was not different between 0 and 4 mg Fe/L from ferrous
lactate, but intake with 8 mg Fe/L from ferrous lactate was reduced. Water intakes of treatments
with 8 mg Fe/L from ferrous lactate, sulfate or chloride were less compared with control water,
but not different among Fe sources. Direct metal analysis commonly used to determine Fe
content of drinking water for livestock greatly underestimates total recoverable Fe (acid-digested
water sample). Total recoverable Fe values as used in our research are listed above.

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................................vi
LIST OF FIGURES ..................................................................................................................viii
LIST OF ABBREVIATIONS ..................................................................................................... x
INTRODUCTION ....................................................................................................................... 1
CHAPTER 1
LITERATURE REVIEW
Introduction........................................................................................................................ 3
Biology of Iron Utilization................................................................................................. 4
Oxidative Stress ................................................................................................................. 8
Iron, Copper, and Zinc Interactions ................................................................................. 10
Iron Chemistry Associated with Biology......................................................................... 12
Iron Microbiology............................................................................................................ 14
Current Drinking Water and Nutritional Recommendations ........................................... 15
Research on Drinking Water with Iron............................................................................ 16
Iron Toxicity in Cattle...................................................................................................... 17
APPENDIX A
FIGURES REFERENCED IN CHAPTER 1
Figures.............................................................................................................................. 23
REFERENCES ............................................................................................................................ 27
CHAPTER 2
2+
3+
EFFECT OF ABOMASALLY DOSED FERROUS (Fe ) OR FERRIC (Fe ) IRON ON
SHORT-TERM IRON STATUS OF LACTATING DAIRY COWS
ABSTRACT................................................................................................................................. 32
INTRODUCTION ....................................................................................................................... 34
MATERIALS AND METHODS................................................................................................. 36
Animals, Experimental Design, and Treatments ............................................................. 36
Preparation of Treatment Solutions ................................................................................. 37
Common treatment administration and sampling (Experiments 1 and 2) ....................... 38
Common Feeding, Water and Milk Data Collection and Analyses................................. 40
Laboratory Analyses ........................................................................................................ 41
Statistical Analyses .......................................................................................................... 44
RESULTS .................................................................................................................................... 46
Treatment by Hour Interactions ....................................................................................... 46
Experiment 1.................................................................................................................... 46
Experiment 2.................................................................................................................... 47
DISCUSSION .............................................................................................................................. 49
Experiment 1.................................................................................................................... 49

iii

Experiment 2.................................................................................................................... 53
Hourly Effects Pooled Across Treatments for Experiments 1 and 2 ............................... 56
CONCLUSIONS ......................................................................................................................... 60
APPENDIX B
FIGURES AND TABLES REFERENCED IN CHAPTER 2
Tables............................................................................................................................... 62
Figures.............................................................................................................................. 70
REFERENCES ............................................................................................................................ 83
CHAPTER 3
2+
3+
EFFECTS OF FERROUS (Fe ) AND FERRIC (Fe ) IRON CONTENT OF DRINKING
WATER ON PREFERENCE AND BEHVIOR OF LACTATING DAIRY COWS
ABSTRACT................................................................................................................................. 87
INTRODUCTION ....................................................................................................................... 89
MATERIALS AND METHODS................................................................................................. 91
Animals, Experimental Design, and Treatments ............................................................. 91
Water Intake Measurement and Behavior Assessment.................................................... 93
Preparation of Treatments................................................................................................ 95
Water Containers ............................................................................................................. 95
Training............................................................................................................................ 95
Common Feeding, Water, and Milk Data Collection and Analyses................................ 96
Statistical Analyses .......................................................................................................... 97
RESULTS AND DISCUSSION .................................................................................................. 99
Treatment Effects on Water Intake .................................................................................. 99
Water Analyses .............................................................................................................. 102
Treatment Effects on Drinking Behavior....................................................................... 103
Correlations between Drinking Behavior and Water Intake.......................................... 104
CONCLUSIONS ....................................................................................................................... 107
APPENDIX C
TABLES REFERENCED IN CHAPTER 3
Tables............................................................................................................................. 109
REFERENCES .......................................................................................................................... 126
CHAPTER 4
CONCLUSIONS AND IMPLICATIONS ............................................................................. 127
REFERENCES .......................................................................................................................... 132
APPENDIX D
ADDITIONAL FIGURES FROM CHAPTER 2 .................................................................. 134
APPENDIX E
DEVELOPMENT OF STUDY PROTOCOL TO MEASURE DRINKING WATER
PREFERENCE AMONG LACTATING DAIRY COWS
ABSTRACT............................................................................................................................... 146
INTRODUCTION ..................................................................................................................... 148
MATERIALS AND METHODS............................................................................................... 149

iv

RESULTS .................................................................................................................................. 155
DISCUSSION ............................................................................................................................ 155
CONCLUSIONS ....................................................................................................................... 158
TABLES .................................................................................................................................... 159
FIGURES................................................................................................................................... 163
REFERENCES .......................................................................................................................... 166

v

LIST OF TABLES

Table 2.1. Pre-experimental dry matter intake, milk yield, water intake and body weight of
cows in Experiments 1 and 2 .......................................................................................62
Table 2.2. Characteristics of treatment infusion solutions in Experiment 1 .................................63
Table 2.3. Characteristics of treatment infusion solutions in Experiment 2 .................................63
Table 2.4. Ingredient and nutrient composition of diets fed in Experiments 1 and 2 ...................64
Table 2.5. Water quality constituent analyses of drinking water in Experiments 1 and 2............65
2+

Table 2.6. Effects of abomasal dosing of different amounts of Fe from ferrous lactate on
blood and liver measurements in Experiment 1...........................................................66
Table 2.7. Pearson correlations coefficients of relationships between independent serum
response variables pooled across treatment and hour for Experiment 1......................67
Table 2.8. Pearson correlations coefficients of relationships between independent liver response
variables pooled across treatment and hour for Experiment 1.....................................67
Table 2.9. Effects of abomasal dosing of ferrous sulfate or ferric sulfate on blood and liver
measurements in Experiment 2....................................................................................68
Table 2.10. Pearson correlations coefficients of relationships between independent serum
response variables pooled across treatment and hour in Experiment 2 ....................69
Table 2.11. Pearson correlations coefficients of relationships between independent liver
response variables pooled across treatment and hour in Experiment 2 ....................69
Table 3.1. Pre-experiment dry matter intake, milk yield and composition, and water intake
of cows in Experiment 1 through 4............................................................................109
Table 3.2. Characteristics of water treatments in Experiment 1 .................................................110
Table 3.3. Characteristics of water treatments in Experiment 2 .................................................110
Table 3.4. Characteristics of water treatments in Experiment 3 .................................................111
Table 3.5. Characteristics of water treatments in Experiment 4 .................................................111
Table 3.6. Ingredient and nutrient composition of diets fed in Experiments 1 through 4...........112

vi

Table 3.7. Water quality constituent analysis of Control drinking water in Experiments 1
through 4 ...................................................................................................................114
Table 3.8. Cumulative water intake (L) by hour of cows in Experiment 1.................................115
Table 3.9. Cumulative water intake (L) by hour of cows in Experiment 2.................................115
Table 3.10. Cumulative water intake (L) by hour of cows in Experiment 3...............................116
Table 3.11. Cumulative water intake (L) by hour of cows in Experiment 4...............................116
Table 3.12. Cumulative drinking behavior durations and frequencies of cows in
Experiment 1 .............................................................................................................117
Table 3.13. Cumulative drinking behavior durations and frequencies of cows in
Experiment 2 .............................................................................................................118
Table 3.14. Cumulative drinking behavior durations and frequencies of cows in
Experiment 3 .............................................................................................................119
Table 3.15. Cumulative drinking behavior durations and frequencies of cows in
Experiment 4 .............................................................................................................120
Table 3.16. Pearson correlation coefficients for the relationships between drinking behavior
and 12 h total water intake in Experiment 1 ............................................................121
Table 3.17. Pearson correlation coefficients for the relationships between drinking behavior
and 12 h total water intake in Experiment 2 .............................................................122
Table 3.18. Pearson correlation coefficients for the relationships between drinking behavior
and 12 h total water intake in Experiment 3 ............................................................123
Table 3.19. Pearson correlation coefficients for the relationships between drinking behavior
and 12 h total water intake in Experiment 4 ............................................................124
Table 5.1. Dry matter intake, milk yield, and water intake of experimental cows .....................159
Table 5.2. Ingredient and nutrient composition of diets fed .......................................................160
Table 5.3. Water quality constituent analyses of Control drinking water ..................................161
Table 5.4. Cumulative water intake (L) of cows pooled for all 3 periods ..................................162
Table 5.5. Cumulative drinking behavior durations and frequencies of cows............................162

vii

LIST OF FIGURES
Figure 1.1. Mechanisms involved in Fe absorption ...................................................................23
Figure 1.2. Molecular mechanisms of Fe utilization in erythrocytes and hepatocytes..............24
Figure 1.3. Formation of heme and Fe-S clusters in erythrocyte mitochondrian ......................25
Figure 2.1. Treatment by hour interaction on hepatic Cu concentrations in Experiment 1 .....70
Figure 2.2. Treatment by hour and pooled hour responses of serum Fe concentration in
Experiment 1 ..........................................................................................................71
Figure 2.3. Treatment by hour and pooled hour responses of serum Fe saturation
concentration in Experiment 1 ...............................................................................72
Figure 2.4. Treatment by hour and pooled hour responses of serum Zn concentration in
Experiment 1 ..........................................................................................................73
Figure 2.5. Treatment by hour and pooled hour responses of hepatic Cu concentration in
Experiment 1 ..........................................................................................................74
Figure 2.6. Treatment by hour and pooled hour responses of hepatic Zn concentration in
Experiment 1 ..........................................................................................................75
Figure 2.7. Treatment by hour and pooled hour responses of serum Fe concentration in
Experiment 2 ..........................................................................................................76
Figure 2.8. Treatment by hour and pooled hour responses of serum unsaturated Fe
binding capacity in Experiment 2 ..........................................................................77
Figure 2.9. Treatment by hour and pooled hour responses of serum Fe saturation
in Experiment 2 .....................................................................................................78
Figure 2.10. Treatment by hour and pooled hour responses of serum Zn concentration in
Experiment 2 .........................................................................................................79
Figure 2.11. Treatment by hour and pooled hour responses of serum α-tocopherol
concentration in Experiment 2 .............................................................................80
Figure 2.12. Treatment by hour and pooled hour responses of whole blood glutathione
peroxidase activity in Experiment 2 .....................................................................81
Figure 4.1 Treatment by hour and pooled hour responses of serum total Fe binding
capacity in Experiment 1 ......................................................................................134

viii

Figure 4.2. Treatment by hour and pooled hour responses of serum total Fe binding
capacity in Experiment 2 ......................................................................................135
Figure 4.3. Treatment by hour and pooled hour responses of serum unsaturated Fe binding
capacity in Experiment 1 ......................................................................................136
Figure 4.4. Treatment by hour and pooled hour responses of serum Cu concentration in
Experiment 1 ........................................................................................................137
Figure 4.5. Treatment by hour and pooled hour responses of serum Cu concentration in
Experiment 2 ........................................................................................................138
Figure 4.6. Treatment by hour and pooled hour responses of serum α-tocopherol
concentration in Experiment 1 .............................................................................139
Figure 4.7. Treatment by hour and pooled hour responses of whole blood glutathione
peroxidase activity in Experiment 1 ....................................................................140
Figure 4.8. Treatment by hour and pooled hour responses of hepatic Fe concentration in
Experiment 1 ........................................................................................................141
Figure 4.9. Treatment by hour and pooled hour responses of hepatic Fe concentration in
Experiment 2 ........................................................................................................142
Figure 4.10. Treatment by hour and pooled hour responses of hepatic Cu concentration in
Experiment 2 .......................................................................................................143
Figure 4.11. Treatment by hour and pooled hour responses of hepatic Zn concentration in
Experiment 2 .......................................................................................................144
Figure 5.1. Treatment by hour response of water intake in period 1 ......................................163
Figure 5.2. Treatment by hour response of water intake across all periods ............................164

ix

ABBREVIATIONS
Ctr1 = copper transporter 1
DMT-1 = divalent metal transporter 1
2+

Fe

3+

Fe

= ferrous iron
= ferric iron

GSH-Px = glutathione peroxidase
Hb = hemoglobin
SEM = standard error of the mean
TDS = total dissolved solids
TIBC = total iron binding capacity
UIBC = unsaturated iron binding capacity

x

INTRODUCTION
Excess Fe in drinking water may have a negative impact on production and health of
lactating dairy cows through decreased water intake, trace mineral interactions, and/or oxidative
stress. Iron supplied by drinking water typically is less than the amount supplied by feed.
However, Fe in water may be more available for absorption than Fe in feed, because ferrous
2+

(Fe ) Fe, the main form of Fe in water (Colter and Mahler, 2006) is more available than ferric
3+

(Fe ) Fe in feed (NRC, 2001). However, there is little research relating Fe concentration of
water with Fe status, health or productivity if lactating dairy cows. In chapter one, literature
related to Fe absorption, metabolism, trace mineral interactions, oxidative stress, and current
information about utilization of Fe in feed and drinking water is reviewed.
High Fe consumption can increase measurements of Fe status; have a potential negative
impact on Cu and Zn status and cause oxidative damage. Iron in water may be more easily
absorbed than Fe in feed, and thus has the potential to cause health problems, though no research
has been completed in dairy cows to prove this theory. Chapter two describes experiments
completed to investigate the effects of abomasally dosing with Fe sources providing differing
concentrations and valences on short-term (12 h) measurements of Fe and antioxidant status and
oxidative stress in lactating dairy cows.
High concentration of Fe in drinking water is a palatability concern for humans (EPA,
2004), and although this has not been shown in lactating dairy cows. The current maximum
tolerable concentration is 0.3 mg Fe/L (EPA, 2004). The third chapter describes four experiments
to evaluate the effects of Fe concentration, valence, and chemical sources (salt form) on drinking
water preference among lactating dairy cows.

1

Chapter 4 provides an overall summary, with conclusions and implications of the
completed research.
We expect our research will provide a foundation for additional research to better
understand the effects of Fe in drinking water on Fe status, health and productivity of lactating
dairy cows.

2

CHAPTER 1
LITERATURE REVIEW

Introduction
There is little information on the effects of Fe concentration and valence in drinking
water on lactating dairy cows (Coup and Campbell, 1964). Most research has been completed in
humans, rats and other non-ruminants. There also is no established upper limit for the
concentration of Fe in drinking water for cattle. The current maximum tolerable concentration
(0.3 mg/L) is based on the human limit for palatability (EPA, 2004). However, it has been
suggested that cattle can tolerate greater concentrations without negative consequences on water
intake or productivity (NRC, 2005).
Anecdotal information suggests that water Fe concentration of 2 mg/L or greater may
negatively impact milk production and cow health (Beede, Michigan State University, East
Lansing, MI, personal communication). It is hypothesized that this effect could be mediated
either through Fe overload and oxidative stress or through drinking water palatability. Iron in
2+

water is presumed to be in the soluble ferrous form (Fe ), which is the form likely absorbed in
the small intestine (Sharp and Srai, 2007). In contrast, Fe from feed (specifically forages) is
3+

present mainly in the ferric form (Fe ) and poorly absorbed (NRC, 2001). Therefore, Fe from
water may have a greater potential to impact milk production and Fe status of lactating cows
more than feed Fe.
Additionally, there is no information regarding the palatability of drinking water with
2+

Fe

3+

or Fe

for lactating dairy cows. Some studies suggest that greater Fe concentration in feed

3

affects the palatability of feed (Hansen et al. 2010; Standish and Ammerman, 1971), though no
conclusions about a relationship between valence and palatability were made definitively.

Biology of Iron Utilization
Iron is an important trace element required by the body for many different aspects of
metabolism. The main function of Fe is as a component of heme-containing proteins in oxygen
transport. Iron is needed in cytochromes and Fe-sulfur proteins for electron transport, and it also
is found in many metalloenzymes and oxygenases used in other metabolic processes (Jacobs,
1980).
Because Fe is vital to oxygen transport, about two-thirds of the total Fe content of the
body is in hemoglobin. A much smaller amount (about 3%) is found in myoglobin, the oxygen
storage molecule in the muscle. There is a small pool of Fe within the cells, stored in vacuoles,
that is used to form Fe-containing proteins (Halliwell and Gutteridge, 1985). The majority of the
remainder is found as ferritin or hemosiderin the storage forms of Fe in various organs, or as
transferrin, the protein used for Fe transport throughout the body (Underwood, 1977).
Absorption. The absorption of Fe in cattle is dependent upon several factors including
age, Fe status and health, secretions of the gastrointestinal tract (acids), and the nutrient
composition of the diet (Underwood, 1977). Pre-ruminants require more absorbed Fe than adult
cattle as a function of body weight; so, by necessity their absorption of Fe is much more efficient
(NRC, 2001). When an animal begins ruminating, its efficiency of absorption decreases (NRC,
2001). Adult cattle do not require additional Fe for growth or increasing blood volume like
calves, so Fe required for normal Fe metabolism is lower. Forages supplied to adult ruminants

4

also contain more than the amount of Fe required, frequently due to soil contamination, although
the availability of this Fe for absorption is unknown. To avoid Fe toxicity, the mechanisms
involved in Fe absorption are down-regulated in adulthood, decreasing the efficiency of
absorption. The NRC (2001) set the absorption coefficient of Fe from common feeds at 0.10.
Absorption efficiency can range from approximately 2 to 20% depending on Fe status, age and
health of the animal.
The mechanisms of Fe absorption are summarized in Figure 1.1. Iron from the diet is
absorbed in the duodenum and the upper portion of the jejunum. Although this information
comes primarily from the human literature, there was no information found that suggests that the
2+

3+

site of absorption is different in ruminants. Iron is absorbed as Fe , and some of the Fe

that is

consumed can be reduced and solubilized by hydrochloric acid secretions in the stomach.
3+

Additionally, Fe

can be reduced by ferrireductases present on the brush border of the

duodenum, likely duodenal cytochrome b (DCYTB; McKie et al., 2001). Ferrous Fe is cotransported into cells with protons by divalent metal transporter 1 (DMT-1; Mackenzie et al.,
2007; Sharp, 2004). Mechanisms of Fe transport within the enterocyte are not well understood,
but depending on Fe status, Fe can either be bound to mucosal ferritin, or brought to the
basolateral surface of the enterocyte for export (Sharp and Srai, 2007). Mucosal ferritin is a
storage compound that remains in the enterocyte until the cell dies and is sloughed off, removing
both the Fe and the cell from the body via fecal excretion (Sharp and Srai, 2007).
When the animal has a need for Fe, instead of being incorporated into mucosal ferritin the
Fe is transported out of the enterocyte by the Fe transporter ferroportin in the basolateral surface
2+

of the cell (Muckenhaler et al., 2008). Once it has reached the blood stream, Fe

5

is oxidized by

the ferroxidase hephaestin, and then bound to transferrin for transport to other areas of the body
(Sharp, 2004). In the case of adequate or excessive Fe, export of Fe from the enterocyte is
another site for regulation. When excess Fe is absorbed, the peptide hormone hepcidin is
released which causes ferroportin to be removed from the membrane and degraded (Nemeth and
Ganz, 2006).
The process of absorption and production and localization of binding proteins involved in
Fe metabolism is regulated through post-transcriptional control by the iron-responsive
element/iron-binding protein (IRE/IRP) system (Muckenthaler et al. 2008). Iron regulatory
proteins (IRP1 and IRP2) bind to IREs on proteins involved in metabolism to allow for Fe
absorption and utilization. When Fe is needed, IRP binds to IRE and stabilizes both DMT1 and
Tfr1 to prevent degradation (Núñez, 2010). Additionally, IRP binding to the IRE region of
ferritin inhibits translation and reduces Fe stored (Núñez, 2010). When a system is overloaded
with Fe, DMT-1 is localized in the cytoplasm, decreasing Fe transport (Trinder et al., 2000). It
was also found that in cells exposed to Fe, approximately 30% of DMT1 on the surface of the
cell is internalized within 10 minutes of Fe exposure (Ma et al., 2006). This suggests that the
internalization of DMT1 operates through a mechanism that is faster than the transcriptional and
translational control of the IRE/IBP system.
Post-absorptive transport and metabolism. The mechanisms of post-absorptive Fe
2+

transport and metabolism are summarized in Figures 1.2 and 1.3. Once absorbed, Fe

is

oxidized by hephastin located in the basolateral membrane. Ferric Fe is bound to the Fe binding
protein transferrin for transport. Most absorbed Fe is brought to erythroid tissue in the bone
marrow for synthesis of hemoglobin. Once transferrin reaches the desired tissue it is bound to

6

transferrin receptors (Tfr1 or Tfr2; Muckenthaler et al., 2008) and the transferrin and receptor
complex are brought into the cell.
Once internalized, Fe is released from transferrin and apotransferrin, or transferrin that is
not bound to Fe is returned to the blood stream. The Fe in the cytosol is brought to the
2+

mitochondria as Fe

to synthesize heme or Fe-S clusters in erythroid precursors (Dunn et al.,

2006), or utilized in phagocytes for the production of lactoferrin, a protein similar to transferrin
that is found in milk (Halliwell and Gutteridge, 1985). Iron not utilized immediately for heme
production is bound to ferritin or hemosiderin for storage (Chua et al., 2007). These compounds
are found mainly in lysosomes of the cells in the liver, spleen and bone marrow (Harrison and
Arosio, 1996). Ferritin is the main storage form, and hemosiderin is a complex very similar to
ferritin (NRC, 2005) although less soluble. Research indicates that ferritin in the lysosomes is
processed through partial degradation of the protein component when Fe is in excess to produce
hemosiderin (Harrison and Arosio, 1996). Ferritin concentrations in the serum were highly
correlated with the total amount of stored Fe in humans (Walters et al., 1973). Miyata and
Furugouri (1987) also examined the relationship between serum ferritin and Fe stored in dairy
cows and found that serum ferritin was highly indicative of Fe nutritional status. This research
group also determined that the suitability of this relationship for measuring Fe status depends on
reproductive stage in dairy cows. Due to the inflammatory processes in the uterus and the
mammary gland that follow parturition, uterine involution and tissue repair, serum ferritin is not
an accurate measure of Fe status until 30 d after parturition (Furugouri et al., 1982).
Excretion and Toxicosis. The natural mechanisms to prevent Fe overload are through
regulation of Fe absorption (Jacobs, 1980) as described above. Therefore, there is no mechanism

7

for Fe excretion to regulate Fe status and only a small amount of Fe is excreted in the feces and
urine (Chua et al., 2007). The tissues most likely to be affected by excess Fe intake and
absorption are the liver, heart and pancreatic beta cells. In dairy cattle, the signs of chronic Fe
toxicosis include reduced feed intake, reduced growth, a decrease in the efficiency of feed
conversion and diarrhea (NRC, 2005).
The body responds quickly to an increase in Fe intake by down-regulating Fe absorption
mechanisms and increasing the amount of Fe stored in ferritin. In rats the synthesis of ferritin in
the liver was maximal 6 h after injection of Fe (Drysdale and Munro, 1966).

Oxidative Stress
Excessive absorbed Fe can cause oxidative stress. During the processes of normal
-

metabolism, reactive oxygen metabolites such as superoxide (O2 ) and hydrogen peroxide
(H2O2) are generated (Miller et al., 1993). These metabolites become a serious concern when
•

they form radical oxygen species, such as the hydroxyl radical (HO ). Iron is a pro-oxidant, and
free metal ions, especially the ferrous ion can create oxygen radicals like the hydroxyl radical
2+

from hydrogen peroxide and superoxide through the Fenton reaction (Fe

3+

+ H2O2  Fe

+

-

OH + OH; Sordillo and Aitkin, 2009). Under normal circumstances, proteins such as ferritin
and transferrin sequester Fe in the body for storage. However, if the mechanisms controlling Fe
absorption fail and excessive Fe is absorbed blood Fe concentration may exceed the capacity of
Fe-binding proteins, leading to free Fe When the reactive oxygen metabolites are formed as a
result of free Fe in the body, a consequence of Fe overload, oxygen radicals are produce. Unless
oxygen radicals are controlled by antioxidant mechanisms apoptosis occurs (Halliwell, 1987).

8

This process is cumulative, because when cells are injured free Fe pools can be disturbed causing
release of the free Fe into the cellular Fe pool, increasing the production of radical oxygen
species (Halliwell, 1987). Halliwell and Gutteridge (1985) found that transferrin, lactoferrin and
ferritin are not generally catalysts of oxygen radical formation; however, they suggested that in
the situation of Fe overload, those proteins also might contribute to radical production.
The formation of hydroxyl radicals is a serious problem in the body. The oxygen radicals
are destructive to lipids, proteins, and DNA (Miller et al., 1993). Radicals are species that
contain one or more unpaired electrons (Halliwell, 1987). Because of this unique structure, these
radicals remove electrons from other stable molecules, causing serious tissue damage. These
damages induce complications in membrane permeability, the function of enzymes, synthesis of
steroid hormones and problems in muscle tone (Miller et al., 1993). The amount of damage and
the concentrations of Fe sufficient to cause peroxidative damage depend on the antioxidant status
of the animal. Older animals are more susceptible than younger animals (Wu et al., 1990).
Some signs of oxidative stress and cellular damage in dairy cows include reduced reproductive
performance, which can be caused by a decrease in uterine contractibility, reduced immune
function, mastitis, parturient paresis, increased incidence of retained placenta and udder edema
(Miller et al., 1993).
Antioxidant status is an important component in the prevention of oxidative stress.
Antioxidants prevent the formation of oxygen radicals and peroxides and prevent oxidative stress
and cellular damage (Sordillo and Aitkin, 2009). Both oxygen radicals and peroxides are
products of normal metabolism, and generally adequate antioxidants are to prevent damage.
Vitamin E, or α-tocopherol, is an antioxidant that reduces radicals (Sordillo and Aitkin, 2009).

9

Superoxide dismutase is an enzyme that reduces the superoxide radical to hydrogen peroxide,
which can then be oxidized by glutathione peroxidase to water and glutathione. However, in the
presence of excess pro-oxidants, like free Fe, antioxidant systems can be overwhelmed and
oxidative stress results. As the load of oxygen radicals increases, the antioxidants are depleted,
leaving the animal in more danger. Calves fed excess Fe (750 mg Fe/kg of ferrous sulfate) had
increased expression of superoxide dismutase mRNA in heart tissue compared to calves fed a
control ration (65 mg Fe/kg), indicating that excess Fe caused oxidative stress (Hansen et al.,
2010).
Excessive Fe in the diet caused damage to the intestinal epithelium, as evidenced by an
increase in flux of labeled mannitol across the intestinal epithelium, and a decrease in epithelial
resistance in the intestinal epithelium of Holstein bull calves fed excessive Fe (750 mg Fe/kg
DM; Hansen et al., 2010). Although the cause of the damage was unclear in this experiment the
authors hypothesized that results pointed toward oxidative stress through lipid peroxidation.

Fe, Cu and Zn Interactions
Iron interactions with other trace elements, specifically Cu and Zn were suspected for
many years. Coup and Campbell (1964) initially noted that hypocupremia developed in cattle
dosed daily with 30 mg Fe/kg body weight from ferric hydroxide (Fe(OH)3) for 7 mo. Continued
research showed that Fe supplementation decreased Cu storage in both cattle and sheep
(Campbell et al., 1974). However, the majority of these studies supplemented Fe using ferrous

10

sulfate (FeSO4), which has antagonistic effects on Cu as well, failing to distinguish the effect of
Fe from the effect of S (Suttle, 1974, Prabowo et al. 1988).
Bremner et al. (1987) demonstrated that it does not take excessive concentrations of Fe in
the diet to have a negative effect on Cu metabolism. They showed that concentrations of Fe
commonly found in some forages are high enough to impact Cu metabolism. Dietary Fe
concentrations of 250, 500, and 750 mg/kg decreased plasma and liver Cu concentrations
without clinical signs in calves fed supplemental Fe as ferrous carbonate (FeCO3). Researchers
in Scotland demonstrated that 800 mg/kg DM of supplemental Fe, supplied as ferrous carbonate,
was sufficient to cause severe Cu deficiency in calves after 16 wk (Humphries et al., 1983). The
supplemented diets reduced hepatic Cu, and decreased plasma Cu, ceruloplasmin, and
erythrocyte superoxide dismutase activity, a Cu and Zn-containing enzyme. Steers supplemented
with 1,000 mg/kg DM of ferrous sulfate had decreased concentrations of Cu in liver and kidney
tissue, and decreased apparent absorption of Cu compared with steers fed a control diet (Standish
et al., 1971).
The most prevalent theory on the interaction between Cu and Fe absorption involves
DMT-1 and competition for transport. Copper is absorbed via metal transporters Ctr1 and DMT1, the main Fe transporter (Sharp, 2004). Additionally, ceruloplasmin, a Cu-containing protein is
a ferroxidase in serum that is essential for changing valence of Fe that affects the ability of Fe to
bind to specific proteins. In the case of Cu deficiency, ceruloplasmin is decreased, which can
decrease Fe status, and result in anemia. Hephaestin also requires Cu, and it is required for the
transport of Fe across intestinal epithelial cells (Sharp, 2004).

11

Steers supplemented with 1,000 mg/kg DM of ferrous sulfate had decreased
concentrations of Zn in the liver compared with steers fed no supplemental Fe (100 mg/kg DM).
(Standish et al., 1971), and as the concentration of ferrous sulfate added to the diet increased (0,
400, and 1,600 mg/kg DM) the concentration of Zn in the liver decreased (Standish et al., 1969).
In contrast, in sheep fed 1,600 mg/kg DM of either ferrous sulfate or ferric citrate
(Fe3(C6H5O7)), there was no change in liver Zn concentration (Standish and Ammerman, 1971).
Supplementing cattle with 1000 mg Fe/kg DM decreased Cu status, but did not impact Zn status
over time (Mullis et al., 2003). Non-heme-Fe inhibits the absorption of Zn when dosed together
2+

in humans, heme-Fe does not (Solomans and Jacob, 1981). Although both Fe
2+

Zn absorption, Fe

3+

and Fe

inhibit

has a greater impact (Solomans et al. 1983).

Unlike the interaction of Fe and Cu, interactions of Fe and Zn are most likely not
mediated through competition at DMT-1 because in a situation of Fe overload, DMT-1
expression and Fe uptake were reduced, but Zn absorption was not impacted in humans (Kordas
and Stolzfus, 2004). Although the mechanism of interaction between these trace elements is
unknown, there is evidence to suggest that the interaction between non-heme-Fe and Zn takes
place in the small intestine (Solomans and Jacob, 1981).
Iron Chemistry Associated with Biology
Ferric iron is the common form in feedstuffs and has a low absorbability. The NRC
(2001) estimated that the absorption coefficient of Fe from the diet is about 0.10 in dairy cattle,
and this number decreases as the amount of Fe consumed increases. A small amount of Fe
2+

reduced to Fe

3+

in the abomasum by hydrochloric acid. In fact, the reduction of Fe

12

3+
2+

to Fe

is
is

3+

important in Fe absorption. Iron is bound to transferrin as Fe , but transferred through the cell
2+

as Fe , to be incorporated into heme or oxidized again to be stored, although the mechanisms
2+

for this are not known (Sharp and Srai, 2007). The ferrous (Fe ) form of Fe is highly available
2+

to the animal. One source estimated that Fe

from ferrous sulfate is 100% available to
3+

mammals (Hurrell, 1997). This form is soluble in water and colorless, and is oxidized to Fe

upon exposure to oxygen, creating ferric oxide (Fe2O3), or rust. Iron toxicity is especially a
concern in areas where the Fe concentration of drinking water is high, because most of the Fe in
water is presumably in the ferrous state (NRC, 2001).
2+

The ratio of Fe

3+

- to - Fe

in water depends on many factors including reduction-

oxidation (redox) potential, pH and other elements and chemical compounds. The oxidation of
2+

Fe

3+

to Fe

increases as both the pH and temperature of water increase (Roekens and Van

Grieken, 1983). However, the rate of oxidation of Fe is only dependent on pH within the range
of 4 to 8. At pH lower than 4 the rate of oxidation is very low, and above a pH of 8, the
2+

oxidation rate is so high that concentrations of Fe

are difficult to detect (Morgan and Lahav,

2007). However, the typical pH of natural groundwater normally falls within the range where
the rate of Fe oxidation is dependent on pH.
Redox potential has a similar effect on the oxidation of Fe as pH. Redox potential is the
tendency of a chemical species to lose or gain electrons. According to Hem and Cropper (1959)
the redox potential of natural water falls between 0.3 to 0.5 volts, and as the redox potential
2+

increases, oxidation of Fe

also increases. However, in a more recent paper Nordstrom (2000)

indicated that natural groundwater is not influenced by redox potential, only pH. In the case of
natural water, redox potential has no practical meaning. In aqueous systems free electrons are

13

very rare, with extremely short half-lives, and are virtually nonexistent in natural groundwater
(Nordstrom, 2000). When considering the redox potential it is important to consider the specific
2+

species, Fe

3+

to Fe

conversion activity for example, and even this is affected by many factors,

such as exposure to oxygen, pH of the solution and dissolved Fe concentration. The redox
2+

potential of Fe /Fe

3+

is difficult to measure, extremely variable (Pierre et al., 2002) and less

influential in Fe oxidation than pH.
3+

Overall, the most important factor in Fe

3+

to Fe

oxidation is elemental oxygen (O2)

content of water, or exposure of the water to oxygen (e.g. in air). This is the reason that aeration
2+

is such a common treatment method for water with high Fe
2+

contact with oxygen, Fe

content. As the water comes in

3+

is oxidized to Fe , and complexes to form ferric oxide or ferric

hydroxide via the Fenton reaction; this precipitates out of solution as a red solid (Stumm and
Lee, 1961). This precipitate can then be filtered out of the water, decreasing the Fe content.

Fe Microbiology
Another problem caused by high Fe concentration in water is the presence of Fe2+

oxidizing bacteria. Iron bacteria are prokaryotes that harvest energy from the oxidation of Fe
3+

to Fe . Although the majority of Fe-oxidizing bacteria exist at very low pH, there are some
bacteria that can oxidize Fe at a near neutral pH, enabling these bacteria thrive in groundwater.
Gallionella ferruginea and Sphaerotilus natans are examples of these Fe-oxidizing bacteria and
they are the species commonly found in water supply pipes (Madigan and Martinko, 2006).
They can be identified by a large amount of Fe precipitate and a red biofilm present in pipes and
water fixtures. Although there is no evidence to suggest that Fe-oxidizing bacteria pose any

14

threat to human or livestock health, they can cause a significant problem with the aesthetic
appeal and palatability of drinking water for humans and reduce water flow and corrode
plumbing equipment.
There is some evidence that excess Fe can contribute to bacterial infections in animals, as
free Fe is required for bacterial growth (NRC, 2001). Baynes et al. (1986) found a correlation
between elevated serum ferritin and acute pneumonia and sepsis. Free Fe also can have a
negative impact on natural resistance to bacterial infection (Bullen et al., 2006).

Current Drinking Water and Nutritional Recommendations
Iron in Drinking Water for Dairy Cattle. Currently the quality recommendation for Fe
content of drinking water for dairy cattle is 0.3 mg/L. This value comes from the EPA (2004)
and is the same as the human limit to avoid health problems and a reduction in palatability. It is
believed that dairy cattle can withstand higher concentrations without affecting water intake or
health (NRC, 2005). However, there is no research to validate this supposition. There is some
anecdotal information suggesting that dairy farmers who installed water treatment systems to
decrease the Fe content of their drinking water observed improvements in cow health and milk
production (Beede, Michigan State University, East Lansing, MI, personal communication).
Some researchers believe that reduced milk yield and health problems from drinking water with
high Fe content may result from reduced water intake (Beede, 2006; Socha et al., 2003).
Socha et al. (2003) estimated that a cow increases Fe intake by 83 mg/d by drinking
water (27.7 gal/d) with an Fe concentration of 0.8 mg/L, which is the average concentration
based on over 3,500 samples taken across the United States. However, this calculation does not

15

consider the potential impact of absorption, which can vary widely depending on age, total Fe
consumption, and other factors. The current belief is that Fe in water may be much more
available than Fe in feeds, due to the difference in valence (Fe

2+

3+

vs. Fe ). Beede (2009)

estimated that a cow drinking 31 gal/d of water with an Fe concentration of only 0.3 mg/L
(assuming 100% absorption) doubles her absorbed Fe intake when considered in addition to her
Fe intake from an average TMR of 20.5 kg DM/d (absorbability of feed Fe assumed to be 0.1).
Yet the research on Fe toxicity from the diet in lactating dairy cows is scarce (Coup and
Campbell, 1964), and research about Fe toxicity from water in lactating dairy cows was not
found in the scientific literature.

Research on Drinking Water with Iron
No research was found testing the effects of Fe in drinking water on normal mammals
with adequate Fe status. However, researchers have investigated the possibility of using water as
a vehicle to deliver Fe to humans and animals with an Fe deficiency. Researchers in Brazil
tested the effects of delivering Fe in water to rats fed an Fe-free diet, and determined that the
inclusion of ferric ammonium citrate, ferrous sulfate or ferrous gluconate increased hemoglobin
concentrations in the treatment rats compared with control rats given deionized water (Ferreira et
al., 1991). This initial evidence of the success of using water to deliver Fe to prevent anemia
resulted in studies in humans in areas of the world that tend to have a problem with anemia.
After the initial experiment with rats, the researchers tested water as a source of Fe in children
with a high incidence of anemia in Brazil (Dutra de Oliveira et al., 1994). They treated the only
water source of a daycare with children from age 2 to 6 yr with ferrous sulfate heptahydrate (20

16

mg/L Fe) for 8 mo. It was estimated that the children drank an average of 500 ml/d, for an
estimated Fe intake of 10 mg/d from the water. Prior to treatment, 58% of the children had
serum hemoglobin concentrations below 11 g/dL, the low end of the normal range. Following
treatment, only 3% of the children had serum hemoglobin concentrations below 11 g/dL, with an
average concentration of 13 g/dL, considered by the investigators to be an acceptable
hemoglobin concentration. They also determined that the serum ferritin concentrations of the
treated children increased from an average before treatment of 13.7 µg/dL to an average after 8
mo of treatment of 25.6 µg/dL. The same research group treated the drinking water of
socioeconomically challenged families in Brazil to contain 10 mg Fe/L of ferrous sulfate and 100
mg/L ascorbic acid (Dutra de Oliveira et al., 1996). The ascorbic acid was added to decrease the
color and turbidity of the treated water and to make it more palatable. They found that both
hemoglobin and serum ferritin were increased after 4 mo, in both children and adults. When
drinking fountains at a daycare in Brazil were supplemented with 10 mg/L Fe and 100 mg/L
ascorbic acid, researchers found that children consuming the water had increased serum
hemoglobin concentrations after 6 mo (Noguiera de Almeida et al., 2005). These children
consumed about 500 ml/d of fortified water, for a total Fe consumption of 5 mg/d from the
water.
No similar studies were found for cattle or other species of domestic livestock.

Iron Toxicity in Cattle
Dietary Fe toxicity in calves was documented by several different studies. In studies in
which 130 d-old male Holstein calves were fed high concentrate diets supplemented with 1,000

17

mg/kg DM of ferrous carbonate, no differences were found between control and treated calves in
the Fe concentration of the liver, kidney, pancreas, spleen, small intestine, muscle, ribs, or bile
(Ho et al., 1984). Miller et al. (1991) reported that in young, pre-weaned heifers fed whole milk
and a starter ration supplemented with 0, 500, 1,000, 2,000 or 4,000 mg/kg DM of ferrous
carbonate (approximately 50% Fe) for 8 wk, weight gains and feed consumption were not
reduced consistently. They found that the lowest average gain and intake were for calves that
were fed ferrous carbonate at 2,000 mg/kg DM though the authors suggested that the difference
in weight gain between 2,000 and 4,000 mg/kg DM did not have a significant meaning. They
also determined that when no milk was fed during the last 2 wk of the trial, both sets of calves
fed 2,000 and 4,000 mg/kg DM had reduced intake and weight gains when compared with the
control calves (0 mg/kg DM). In contrast, Jenkins and Hidiroglou (1987) found no differences
for calves fed milk replacer with supplemental ferrous sulfate heptahydrate at concentrations of
100, 500, 1,000 or 2,000 mg Fe/kg DM, but calves on the 5,000 mg Fe /kg DM treatment had
lower DMI and average daily gain. Also, when compared with calves on the lowest dietary
treatment, calves on the 5,000 mg Fe/kg DM treatment had increased non-heme Fe, ferritin and
hemosiderin in the liver, kidney and spleen. Also, as milk replacer Fe concentration increased,
there was a concurrent increase in Fe concentrations in blood plasma, bile, gall bladder, muscle,
heart, spleen, liver and kidney.
Few studies have investigated the effects of high dietary Fe in adult ruminants. One
study in Florida tested the effects of ferrous sulfate in beef steers (Standish et al., 1969).
Researchers fed diets containing 0, 400 or 1,600 mg Fe/kg DM supplemental ferrous sulfate, and
investigated the effects on average daily gain, feed-to-gain ratio, blood measurements and the

18

mineral content of various body tissues. They found that as the dietary Fe concentration
increased, the amount of feed consumed by the steers decreased, as did the average daily gain.
However, there was no difference between treatments in hemoglobin or hematocrit
concentrations of the blood plasma. The Fe contents of the liver, spleen, kidney and heart were
increased as dietary Fe increased. They also found an inverse relationship between the Fe content
and the Cu and Zn content of the liver. As the dietary Fe concentration increased the Cu and Zn
content of the liver decreased. When 1,600 mg Fe/kg DM of either ferrous sulfate or ferric
citrate was fed to growing wethers, hematocrit decreased by the Fe treatments when compared
with control, and hemoglobin was numerically lower (Standish and Ammerman, 1971).
However, in agreement with Standish et al., (1969), the Fe concentrations of the liver and the
spleen were greater in the wethers supplied excess dietary Fe.
Although most research on Fe toxicity of dairy cattle was completed in pre-ruminating
dairy calves, one New Zealand study investigated the effects of pasture heavily irrigated with
water with a high Fe content on non-lactating dairy cattle (Coup and Campbell, 1964). Iron
content of the pasture available to the Fe treatment group increased through periodic irrigation,
reaching concentrations ranging from 570 to 9,980 mg/kg (dry basis). Researchers were not
positive whether the Fe was taken up by the grass, or was simply on the surface. This was
achieved through irrigating the pasture with water with a high Fe concentration (17 mg/L) three
times through the two-month research period for a total of 68 cm of water applied on the high Fe
pasture. Pasture of the control group was not irrigated with the high Fe water. There was a loss of
live weight in both groups, but animals in the control group, on pasture that never had Fe content
above 420 mg/kg DM, lost less live body weight, and lost weight less rapidly than the cows on

19

high Fe pasture. They also found that the cows on high Fe pasture continuously had loose,
abnormally colored manure (Coup and Campbell, 1964). In a separate experiment, the same
3+

researchers tested the effects of dosing a lactating dairy cow on pasture with 130 g of Fe

per

day for 5 d in the form of ferric chloride. The dosed cow’s (only one cow) milk yield decreased
rapidly, as did her live body weight. Her feces also were dark in color, loose and malodorous
(Coup and Campbell, 1964). Cattle dosed for 10 d twice daily with increasing amounts of ferric
hydroxide, (0, 15, 30 and 60 g/d or approximately, 0, 7.5, 15 and 30 g Fe/d) had a dose-related
decrease in milk and fat yield, body weight and reduced feed conversion efficiency (Coup and
Campbell, 1964). No other research was found that tested the effects of high amounts of dietary
Fe on lactating dairy cattle, nor the effects of high Fe in drinking water.
Therefore, because research on the effects of dietary Fe in lactating dairy cows is scant
and with variable results and research on the effects of drinking water with high Fe
concentrations is very limited, and non-existent in the case of lactating dairy cows, we have
2+

many questions. The only knowledge that we currently have about the effects of Fe

from

drinking water on lactating dairy cows is through personal accounts (Beede, Michigan State
University, East Lansing, MI, personal communication). The research that supports the current
guidance for Fe content of drinking water for cattle is based on recommendations for humans,
although it is suggested that cattle can tolerate higher concentrations (NRC, 2005). Overall, we
hypothesize that indicators of Fe status will be impacted short-term when lactating dairy cows
are dosed abomasally with Fe in a water solution. We also hypothesize that the preference of
cows for drinking water will be affected by the concentration and valance of Fe in the water.
Through two separate approaches we intend to characterize the lactating dairy cow’s

20

physiological response and preference for Fe at different concentrations, of different valences
2+

(Fe

3+

vs. Fe ) and from different chemical salts delivered through drinking water.

21

APPENDIX A
FIGURES REFERENCED IN CHAPTER 1

22

FIGURES
Figure 1.1. Mechanisms involved in Fe absorption

Apical
Surface !!

Basolateral
Surface

binding protein

transferrin-like
protein
transferrin
hepcidin

Fe3+

Fe2!" DMT#1

Fe3!"

hephaestin

ferrireductase

Fe2!"

Fe2!"
lysosome

Intestinal
Lumen

ferroportin

mucosal
ferritin

Duodenal Enterocyte

3+

Plasma

2+

Dietary ferric iron (Fe ) is reduced by a ferrireductase to ferrous iron (Fe ) in the intestinal
2+

lumen. The Fe

binds to a binding protein and this complex binds to divalent metal transporter

1 (DMT-1) which is then internalized. If Fe status and stores are high, Fe is bound to mucosal
ferritin, which is stored in the lysosome until the enterocyte is sloughed off and excreted in feces.
If Fe status and body stores are inadequate, Fe travels through the enterocyte through unknown
mechanisms, and exported from the enterocyte via the basolateral transporter ferroportin. Once
in the bloodstream, Fe

2+

3+

is oxidized to Fe

by hephaestin and bound to transferrin for transport.

In a situation of Fe overload hepcidin is released, which causes removal of ferroportin from the
basolateral surface and degradation of ferroportin. [Adapted from Beard and Dawson (1997) and
Sharp (2007)]. For interpretation of the references to color in this and all other figures, the reader
is referred to the electronic version of this thesis.

23

Figure 1.2. Molecular mechanisms of Fe utilization in erythroid pre-cursers and hepatocytes

transferrin
endosome

transferrin receptor
ferrireductase

transferrin

Fe3+!!

transferrin receptor

Fe2+!!
DMT-1

on
ch
to
mi

a
dri

Fe2+ !!
ferritin
Fe2+!!

a.) Erythroid Precursor

Fe3+!!

ceruloplasmin !!
DMT-1
Fe3+!!

transferrin

Fe2"!

endosome

Fe2+ !!
ferroportin
b.) Hepatocyte

In both pre-erythrocytes and hepatocytes, transferrin is bound to a transferrin receptor (Trf1 or
Trf2) on the cell plasma membrane and internalized through endocytosis into the endosome. The
3+

low pH in the endosome causes the release of Fe from transferrin. Ferrous iron (Fe ) is then
2+

reduced to ferrous iron (Fe ) by a ferrireductase, and exported from the endosome by DMT-1.
a.) In erythroid precursors after release of Fe from the endosome, Fe is transported to the
mitochondria for heme and Fe-S cluster synthesis. b.) In the hepatocyte Fe is bound to ferritin for
2+

storage, or can be exported as Fe

3+

via ferroportin, oxidized to Fe

by ceruloplasmin and bound

to transferrin again for transport. [Adapted from Chua et al. (2007) and Dunn et al. (2006)].

24

Figure 1.3. Formation of heme and Fe-S clusters in erythroid pre-curser mitochondrion

Fe2!"
mitoferrin

as
at
el
ch

rro
fe

Fe-S clusters

Fe2!"

e

Fe ""S
heme
mitochondrial ferritin

ABCB7

heme

Erythroid Precursor
Mitochondrion
Fe ""S

transporter
heme

2+

Ferrous iron (Fe ) is transported into the mitochondria via a transporter, possibly mitoferrin.
The path of Fe inside the mitochondria is possibly determined by frataxin. The Fe can be used to
synthesize heme by ferrochelatase, it can be sequestered in mitochondrial ferritin, or it can be
used for Fe-S synthesis. Fe-S clusters are exported through the transporter ABCB7. Heme is
transported from the mitochondria through a heme transporter (possibly ABCG2, FLVCR or
ABC-me) [Adapted from Dunn et al. (2006)].

25

REFERENCES

26

REFERENCES
Baynes, R., W. Bezwoda, T. Bothwell, Q. Khan, and N. Mansoor. 1986. The non-immune
inflammatory response: serial changes in plasma iron, iron-binding capacity, lactoferrin,
ferritin and C-reactive protein. Scand. J. Clin. Invest. 46:695-704.
Beard, J. L., and H. D. Dawson. 1997. Iron. Pages 275-334 in Handbook of Nutritionally
Essential Mineral Elements. B. L. O’Dell and R. A. Sunde, ed. Marcel Dekker, New
York, NY.
Beede, D. K. 2006. Evaluation of water quality and nutrition for dairy cattle. High Plains Dairy
Conf. Proc., pp. 129-154
Beede, D. K. 2009. Solving bad water problems for thirsty cows. Western Dairy Management
Conf. Proc. pp., 1-9.
Bremner, I., W. R. Humphries, M. Phillippo, M. J. Walker, and P. C. Morrice. 1987. Ironinduced copper deficiency in calves: dose-response relationships and interactions with
molybdenum and sulphur. Anim. Prod. 45:403-414.
Bullen, J. J., H. J. Rogers, P. B. Spalding, and C. G. Ward. 2006. Natural resistance, iron and
infection: a challenge for clinical medicine. J. Med. Microbiol. 55:251-258.
Campbell, A. G., M. R. Coup, W. H. Bishop, and D. E. Wright. 1974. Effect of elevated iron
intake on the copper status of grazing cattle. N. Z. J. Agric. Res. 17:393-399.
Chua, A. C. G., R. M. Graham, D. Trinder, and J. K. Olynyk. 2007. The regulation of cellular
iron metabolism. Crit. Rev. Clin. Lab. Sci. 44:413-459.
Colter, A., and R.L. Mahler. 2006. Iron in Drinking Water. PNW 589. Pacific Northwest
Extension Publication.
Coup, M. R., and A. G. Campbell. 1964. The effect of excessive iron intake upon the health and
production of dairy cows. N. Z. J. Agric. Res. 7: 624-638.
Drysdale, J. W., and H. N. Munro. 1966. Regulation of synthesis and turnover of ferritin in rat
liver. J. Biol. Chem. 241:3630-3637.
Dunn, L. L., Y. S. Rahmanto, and D. R. Richardson. 2006. Iron uptake and metabolism in the
new millennium. Trends in Cell Biology. 17:93-100.
Dutra-de-Oliveira, J. E., J. B. Ferreira, V. P. Vasaconcellos, and J. S. Marchini. 1994. Drinking
water as an iron carrier to control anemia in preschool children in a day-care center. J.
Am. Coll. Nutr. 13(2):198-202.
Dutra-de-Oliveira, J. E., M. M. Amaral Scheid, I. D. Desai, and S. Marchini. 1996. Iron
fortification of domestic drinking water to prevent anemia amon glow socioeconomic
families in Brazil. Int. J. Food Sci. Nutr. 47:213-219.
27

EPA (U.S. Environmental Protection Agency). 2004. 2004 Edition of the drinking water
standards and health advisories. Available at http://water.epa.gov/action/
advisories/drinking/upload/2009_04_27_criteria_drinking_dwstandards2004.pdf.
Accessed October 23, 2009.
Ferreira, J. F., R. A. Aranda, M. P. Bianchi, I. D. Desai, and K. E. Dutra de Oliveira. 1991.
Utilization of drinking water as a vehicle for nutrients: experimental studies with iron.
Arch. Latinoam. Nutr. 41(3):400-408.
Furugouri, K., Y. Miyata, and K. Shijimaya. 1982. Ferritin in the blood serum of dairy cows. J.
Dairy Sci. 65:1529-1534.
Halliwell, B. 1987. Oxidants and human disease: some new concepts. FASEB J. 1:358-364.
Halliwell, B., and J. M. Gutteridge. 1985. The importance of free radicals and catalytic metal
ions in human diseases. Mol. Aspects Med. 8(2):89-193
Hansen, S. L., M. S. Ashwell, A. J. Moeser, R. S. Fry, M. D. Knutson, and J. W. Spears. 2010.
High dietary iron reduces transporters involved in iron and manganese metabolism and
increases intestinal permeability in calves. J. Dairy Sci. 93:656-665.
Harrison, P. M., and P. Arosio. 1996. The ferrition: molecular properties, iron storage function
and cellular regulation. Biochim. Biophys. Acta. 1275(3):161-203.
Hem, J. D., and W. H. Cropper. 1959. Chemistry of iron in natural water: survey of ferrous-ferric
chemical equilibria and redox potentials. U. S. Geol. Surv. Water-Supply Pap. 1459-A.
Ho, S. Y., W. J. Miller, R. P. Gentry, M. W. Neathery, and D. M. Blackmon. 1984. Effects of
high but nontoxic dietary manganese and iron on their metabolism by calves. J. Dairy
Sci. 67:1489-1495.
Huebers, H. A., E. Huebers, E. Csiba, W. Rummel, and C. A. Finch. 1993. The significance of
transferrin for intestinal absorption. Blood. 61:283-290.
Humphries, W. R., M. Phillipo, B. W. Young, and I. Bremner. 1983. The influence of dietary
iron and molybdenum on copper metabolism in calves. Br. J. Nutr. 49:77-86.
Hurrell, R. F. 1997. Preventing iron deficiency through food fortification. Nutr. Rev. 55(6):210222.
Jacobs, A. 1980. The pathology of iron overload. Pages 427-459 in Iron in Biochemistry and
Medicine. Vol. 2. A. Jacobs and M. Worwood, ed. Academic Press, London, UK.
Jenkins, K. J., and M. Hidiroglou. 1987. Effects of excess iron in milk replacer on calf
performance. J. Dairy Sci. 70:2349-2354.
Kordas, K., and R. J. Stolzfus. 2004. New evidence of iron and zinc interplay at the enterocyte
and neural tissues. J. Nutr. 134:1295-1298.

28

Ma, Y., M. Yeh, K. Yeh, and J. Glass. 2006. Iron imports. V. Transport of iron through the
intestinal epithelium. Am. J. Physiol. Gastrointest. Liver Physiology. 290:G417-G422.
Mackenzie, B., Takanga, H., N. Hubert, A. Rolfs, and M. A. Hediger. 2007 Functional properties
of multiple isoforms of human divalent metal-iron transporter 1 (DMT1). Biometals.
16:41-54.
Madigan, M., and J. Martinko. 2006. Metabolic Diversity. Pages 531-592. In Brock Biology of
Microorganisms. 11th ed. Pearson Prentice Hall, Upper Saddle River, NJ.
McKie, A. T., D. Barrow, G. O. Latunde-Dada, A. Rolfs, G. Sager, E. Mudaly, M. Mudaly, C.
Richardson, D. Barlow, A. Bomford, T. J. Peters, K. B. Raja, S. Shirali, M. A. Hediger,
F. Farzaneh, and R. J. Simpson. 2001. An iron-regulated ferric reductase associated with
the absorption of dietary iron. Science. 291:1755-1759.
Miller, J. K., E. Brzezinska-Slebodzinska, and F. C. Madsen. 1993. Oxidative stress,
antioxidants, and animal function. J. Dairy Sci. 76:2812-2823.
Miller, W. J., R. P. Gentry, D. M. Blackmon, and H. H. Fosgate. Effects of high dietary iron as
ferrous carbonate on performance of young dairy calves. J. Dairy Sci. 74:1963-1967.
Miyata, Y., and K. Furugouri. 1987. The relationship between serum ferritin concentration and
tissue non-heme iron or tissue ferritin in dairy cattle. Nippon Juigaku Zasshi. 49:11571159.
Morgan, B., and O. Lahav. 2007. The effect of pH on the kinetics of spontaneous Fe(II)
oxidation by O2 in aqeous solution – basic principles and a simple heuristic description.
Chemosphere. 68:2080-2084.
Muckenthaler, M. U., B. Galy, and M. W. Hentze. 2008. Systemic iron homeostasis and the ironresponsive element/iron-regulatory protein (IRE/IBP) regulatory network. Ann. Rev.
Nutr. 28:197-213.
Mullis, L. A., J. W. Spears, and R. L. McCraw. 2003. Effects of breed (Angus vs Simmental) and
copper and zinc source on mineral status of steers fed high dietary iron. J. Anim. Sci. 81:
318-322.
National Research Council. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad.
Sci., Washington, DC.
National Research Council. 2005. Mineral Tolerance of Animals. 2nd rev. ed. Natl. Acad. Sci.,
Washington, DC.
Nemeth, E., and T. Ganz. 2006. Regulation of iron metabolism by hepcidin. Ann. Rev. Nutr.
26:323-342.
Noguiera de Almeida, C. A., J. E. Dutra-de-Oliveira, G. C. Crott, A. Cantolini, R. G. Ricco, L.
A. Del Ciampo, and M. E. Costa Baptista. 2005. Effect of fortification of drinking water

29

with iron plus ascorbic acid or with ascorbic acid alone on hemoglobin values and
anthropometric indicators in preschool children in day-care centers in Southeast Brazil.
Food Nutr. Bull. 26:259-265.
Nordstrom, D.K. 2000. Aqueous redox chemistry and the behavior of iron in acid mine waters. p.
43–47. In R.T. Wilkin et al. (ed.) Worksh. in Monitoring Oxidation–Reduction Processes
for Ground-water Restoration, Dallas, TX. 25–27 Apr. 2000. Worksh. Summary. USEPA
Natl. Risk Manage. Res. Lab., Cincinnati, OH.
Núñez, M. T. 2010. Regulatory mechanisms of intestinal iron absorption – Uncovering of a fastresponse mechanism based on DMT1 and ferroportin endocytosis. Biofactors. 36:88-97.
Prabowo, A., J. W. Spears, and L. Goode. 1988. Effects of dietary iron on performance and
mineral utilization in lambs fed a forage-based diet. J. Anim. Sci. 66:2028-2035.
Pierre, J. L., M. Fontecave, and R. R. Crichton. 2002. Chemistry for an essential biological
process: the reduction of ferric iron. Biometals. 15:341-346.
Roekens, E. J., and R. E. Van Grieken. 1983. Kinetics of iron(II) oxidation in seawater of various
pH. Marine Chemistry. 13:195-202.
Sharp, P. 2004. The molecular basis of copper and iron interactions. Proc. Nutr. Soc. 63:563-569.
Sharp, P., and S. K. Srai. 2007. Molecular mechanisms involved in intestinal iron absorption.
World. J. Gastroenterol. 13(35):4716-4724.
Socha, M. T., S. M. Ensley, D. J. Tomlinson, and A. B. Johnson. 2003. Variability of water
composition and potential impact on animal performance. Proceedings of the 2003
Intermountain Nutrition Conference, pp. 85-96.
Solomans, N. W., and R. A. Jacob. 1981. Studies on the bioavailability of zinc in humans: effects
of heme and nonheme iron on the absorption of zinc. Am. J. Clin. Nutr. 34:475-482.
Solomans, N. W., O. Pineda, F. Viteri, and H. H. Sandstead. 1983. Studies on the bioavailability
of zinc in humans: mechanism of the intestinal interaction of nonheme iron and zinc.
113:337-349.
Sordillo, L. M., and S. L. Aitken. 2009. Impact of oxidative stress on the health and immune
function of dairy cattle. Vet. Immunol. Immunopathol. 128:104-109.
Standish, J. F., C. B. Ammerman, C. F. Simpson, F. C. Neal, and A. Z. Palmer. 1969. Influence
of graded levels of dietary iron as ferrous sulfate on performance and tissue mineral
composition of steers. J. Anim. Sci. 29:496-503.
Standish, J. F., and C. B. Ammerman. 1971. Effect of excess dietary iron as ferrous sulfate and
ferric citrate on tissue mineral composition of sheep. J. Anim. Sci. 33:481-484.

30

Standish, J. F., C. B. Ammerman, A. Z. Palmer, and C. F. Simpson. 1971. Influence of dietary
iron and phosphorus on performance, tissue mineral composition and mineral absorption
on steers. J. Anim. Sci. 33:171-178.
Stumm, W., and G. F. Lee. 1961. Oxygenation of ferrous iron. Ind. Eng. Chem. 53:143-146.
Suttle, N. F. 1974. Effects of organic and inorganic sulphur on the availability of dietary copper
to sheep. Br. J. Nutr. 32:559.
Trinder, D., P. S. Oates, C. Thomas, J. Sadleir, and E. H. Morgan. 2000. Localisation of divalent
metal transporter 1 (DMT1) to the microvillus membrane of rat duodenal enterocytes in
iron deficiency, but to hepatocytes in iron overload. Gut. 46:270-276.
Underwood, E. J. 1977. Iron. Pages 13-55 in Trace Elements in Human and Animal Nutrition.
4th ed. Academic Press. New York, NY.
Walters, G. O., F. M. Miller, and M. Worwood. 1973. Serum ferritin concentration and iron
stores in normal subjects. J. Clin. Pathol. 26:770-772.
Wu, W., M. Meydani, S. N. Meydani, P. M. Burklund, J. B. Blumberg, and H. N. Munro. 1990.
Effect of dietary iron overload on lipid peroxidation, prostaglandin synthesis and
lymphocyte proliferation in young and old rats. J. Nutr. 120:280-289.

31

CHAPTER 2
2+

3+

EFFECTS OF ABOMASALLY DOSED FERROUS (Fe ) OR FERRIC (Fe ) IRON ON
SHORT-TERM IRON STATUS OF LACTATING DAIRY COWS
ABSTRACT
The majority of Fe naturally occurring in drinking water drawn from underground wells
2+

is in the more bioavailable ferrous (Fe ) valence. In contrast, the majority of Fe in feeds is of
3+

the ferric (Fe ) valence, and poorly absorbed by ruminants. Therefore, Fe from drinking water,
though present in lower concentrations, could have a greater impact on Fe status, potential Fe
toxicity and oxidative stress than feed Fe. Our objective was to evaluate the short-term effects of
abomasally infused Fe

2+

2+

3+

and the difference in short-term effects of Fe and Fe

when

administered at concentrations to simulate total daily Fe intake from high-Fe drinking water. In
each experiment, 6 mid-lactation Holstein cows were assigned in a replicated 3 x 3 Latin Square
balanced for treatment sequences. Treatments in Experiment 1 were 0, 0.75 and 1.5 mg of Fe
from ferrous lactate [Fe(C3H5O3)2] per kg body weight (BW). Treatments were calculated to
approximate 0, 4.5 and 9 mg Fe/L concentrations in drinking water, respectively, and were dosed
directly into the absomasum in 1 L of deionized water. Experiment 2 differed only in the
treatments administered: 0 mg Fe and 1.5 mg Fe/kg BW of either ferrous (FeSO4) or ferric
sulfate [Fe2(SO4)3]. In each experiment blood samples were taken at 0 h and hourly for 12 h
post-dosing, and liver biopsies were taken at 0, 18 and 36 h post-dosing. There were no effects of
either concentration or valence of Fe dosed on serum Fe, unsaturated Fe binding capacity,
percent Fe saturation, or Zn. In Experiment 1, cows infused with 0.75 mg/kg BW of ferrous
lactate had increased serum total Fe binding capacity and hepatic Cu concentrations. Cows

32

infused with 1.5 mg Fe/kg BW had increased serum α-tocopherol concentrations. In Experiment
2, there was a tendency for cows administered ferric sulfate to have lower serum Fe than those
infused with no Fe or ferrous sulfate. Cows infused with ferrous sulfate had lower serum Cu
concentrations than those in other treatments. Treatments did not influence short-term Fe status,
suggesting that it may take chronic exposure to Fe in drinking water, or greater Fe concentrations
to impact Fe status and productivity of lactating cows.

33

INTRODUCTION
There is very little research describing non-dietary Fe toxicity in dairy cows. Lactating
cows dosed intraruminally with 130 g Fe/d as ferric hydroxide [Fe(OH)3] for 10 d presented
clinical signs of extreme toxicity (Coup and Campbell, 1964). Excess dietary Fe can decrease
weight gain and feed consumption in young ruminants and steers (Miller et al., 1991; Jenkins
and Hidiroglou, 1987; Standish et al. 1969) apparently through Fe overload and oxidative stress.
These same mechanisms may impact dairy cows consuming drinking water with high
concentrations of Fe. Even though the concentration of Fe in most drinking water and amount
consumed is much less than that from a typical dairy cow ration, Fe in water may be ferrous iron
2+

(Fe ) (Colter and Mahler, 2006). Ferrous iron is more absorbable than Fe from the ration,
3+

which is typically in the ferric form (Fe ) (NRC, 2001).
Iron metabolism is regulated through absorption, as very little Fe is excreted. Pathways of
Fe absorption are down regulated in response to high Fe concentration from feed or water
(Jacobs, 1980). As intake of Fe increases, Fe receptor divalent metal transporter 1 (DMT-1) on
the apical surface of the duodenal enterocyte is internalized (Sharp, 2004). Additionally,
hepcidin, a peptide hormone involved in Fe homeostasis, is released which inhibits the
basolateral Fe transporter ferroportin and reduces Fe export from the enterocyte. Despite tight
regulation, the mechanisms can be overwhelmed in cases of Fe overload.
Excess Fe also can lead to oxidative stress. Iron is a pro-oxidant (Sordillo and Aitken,
2009) and when Fe homeostasis is achieved, absorbed Fe is bound to Fe-binding proteins such as
transferrin or sequestered in ferritin. However, when excess Fe is consumed and reaches the

34

lumen of the small intestine, high concentrations can damage the intestinal mucosa. Although
most Fe is absorbed and incorporated into mucosal ferritin, and the rest excreted in feces, excess
Fe can overwhelm absorptive mechanisms. Dairy calves exposed to high dietary Fe
concentrations (750 mg/day of FeSO4 for 56 d) had increased intestinal permeability (Hansen et
al., 2010).
Multiple studies have used solutions of Fe in water to improve the Fe status of Fedeficient rats and humans (Ferreira et al., 1991; Dutra de Oliveira et al., 1994). After determining
that the inclusion of Fe
2+

et al., 1991), Fe

2+

in drinking water increased hemoglobin concentration in rats (Ferreira

was supplied in drinking water to children with low Fe status, for an estimated

total intake of 10 mg Fe/d. After 8 mo of supplementation, the percentage of children with
deficient concentrations of hemoglobin decreased from 58 to 3% (Dutra de Oliveira et al., 1994).
2+

This evidence suggests that low concentrations of Fe , delivered through water can impact
measurements of Fe status.
Therefore, we hypothesize that the majority of Fe, naturally occurring in drinking water
2+

as Fe

could increase measurements of Fe status and potentially cause toxicity if present in

great enough concentrations. Our objective was to evaluate the short-term effects of amount and
valence of abomasally infused Fe on the Fe status of mid-lactation dairy cows. We expected that
as the amount of Fe dosed increases, measurements of Fe status would increase, and Fe
3+

increase measurements of Fe status more than Fe .

35

2+

would

MATERIALS AND METHODS
The Institutional Animal Care and Use Committee at Michigan State University approved
all experimental procedures (AUF # 03/10-024-00).

Animals, Experimental Design, and Treatments
Experiment 1. Six ruminally fistulated Holstein cows [219 ± 85 DIM (days in milk)]
were assigned randomly in a replicated 3 x 3 Latin Square design involving 3 periods and 3
experimental treatments. There was a preliminary adjustment period of 2 wk prior to the first
experimental period and 7 d between the experimental periods. Treatments were 0 (0Fe), 0.75
2+

(0.75Fe) and 1.5 (1.5Fe) mg Fe /kg BW from ferrous lactate [Fe(C3H5O3)2]. Treatments were
made iso-lactate through lactic acid addition. Each treatment was administered as a single dose
in 1 L of distilled, deionized water into the abomasum (Table 2.2). The 0.75Fe and 1.5Fe
treatments were formulated to contain the total amount of Fe that a 600-kg lactating cow
drinking 100 L of water/d would consume in a single day if drinking water with total Fe
concentrations of 4.5 and 9 mg/L, from water according to total recoverable analysis (EPA
method 200.7, 1994), or 1, and 2.25 mg/L from water according to direct Fe analysis (EPA
method 200.7, 1994), respectively. One animal was removed from the experiment during period
2 due to illness, but recovered, and returned for period 3.
Experiment 2. Six ruminally fistulated Holstein cows (205 ± 22 DIM) were assigned
randomly in a replicated 3 x 3 Latin Square design involving 3 experimental treatments and 3
periods. There was an adjustment period of 2 wk prior to the first experimental period and 7 d

36

between the experimental periods. Experimental treatments were 0 (Control) and 1.5 mg Fe/ kg
BW from either ferrous (ferrous sulfate; FeSO4) or ferric sulfate (ferric sulfate; Fe2(SO4)3.
Treatments were made iso-sulfate through sulfuric acid addition. Each treatment was
administered as a single dose in 1.5 L of distilled, deionized water into the abomasum (Table
2.3). The ferrous sulfate and ferric sulfate treatments were formulated to contain approximately
the total amount of Fe that a 600-kg lactating cow drinking 100 L of water/d would consume if
drinking water with total Fe concentration of 9 mg/L, from water according to total recoverable
Fe analysis (EPA method 200.7, 1994), or 2.25 mg/L from water according to direct Fe analysis
(EPA method 200.7, 1994), respectively.

Preparation of Treatment Solutions.
Experiment 1. Cows were weighed weekly beginning 2 wk before period 1 to determine
the amount of Fe each cow would receive, administered as a function of BW. Treatment
solutions of 0, 0.75 or 1.5 mg Fe/kg BW from ferrous lactate were prepared by adding ferrous
lactate to 125 mL of distilled, deionized water, heating and stirring well until the Fe compound
dissolved. The amount of lactic acid required to make treatments iso-lactate was added at this
time. This solution was diluted with 875 ml of distilled, deionized water to bring the total
treatment solution to 1 L. The pH of the solution was measured and recorded (Table 2.2).
Experiment 2. Cows were weighed weekly beginning 2 wk before period 1 to determine
the amount of Fe each cow would receive, administered as a function of BW. Treatment
solutions of 0, 1.5 mg Fe/kg BW from ferrous or ferric sulfate were prepared by adding the

37

appropriate Fe compound to 200 mL of distilled deionized water, heating and stirring well until
the Fe compound dissolved into solution. The amount of sulfuric acid required to make
treatments iso-sulfate was added at this time. This solution was diluted with 1,350 mL of
distilled, deionized water to bring the total solution volume to 1.5 L. The pH of the solution was
measured and recorded (Table 2.3).

Common Treatment Administration and Sampling (Experiments 1 and 2).
Experimental periods were 36 h in length. Treatment solutions were administered as a
single dose in approximately 1 min through a polyvinyl tube passed through the ruminal fistula
and the reticulo-omasal orifice into the proximal abomasum. An insertion device, similar to that
described by Gressley et al. (2006), was crafted from a 2.5 cm (inside diameter) polyvinyl
chloride pipe (schedule 40 PVC), cut to approximately 23 cm in length and smoothed with
sandpaper. One end was sanded into a smooth beveled edge to enable easy insertion into the
reticulo-omasal orifice. Once the insertion device was placed into the abomasum, a dosing tube
(1.58 cm outside diameter, 1.27 cm inside diameter) was inserted through the insertion device so
that approximately 40 cm of the tube was located through the reticulo-omasal orifice beyond the
rumen. Five hundred mL of water was initially poured via funnel through the dosing tube to
confirm placement of the tube in the abomasum. After confirmation, 1 L treatment solution was
poured into the abomasum. After all of the treatment solution was administered, 500 ml of
additional water was administered through the tube into the abomasum to purge any remaining

38

treatment solution. In Experiment 2, 250 ml of water was used to confirm initial placement of
the stomach tube in the abomasum, and the treatment solution was 1.5 L.
Jugular vein catheters were placed at least 24 h before the beginning of each experimental
period in the contralateral vein from the previous period and removed at the end of each
experimental period. Catheters were in place for approximately 75 h during each experimental
period and all catheters were maintained twice daily by checking for patency and flushing with
Na-heparin (100 IU/ml heparin in 0.9% sterile saline solution). Blood samples from indwelling
jugular catheters were taken hourly -6, -5, -4, -3, -2, -1, 0 (just before treatment dosing), and
hourly for the next 12 h. Pre-treatment blood sampling began at 1100 h, and continued until
1600 h. Cows were milked in the Dairy Teaching and Research Center (Michigan State
University, East Lansing, MI) parlor at 1700 h, and treatment administration was completed at
1800 h. Orts were collected at 1000 h and feeding occurred at the same time (1115 h) each day in
each period. Blood samples were allowed to clot and then spun in a centrifuge (1,000 g).
Serum was harvested, transferred to microcentrifuge vials and stored at -10°C until analyses for
Fe, Cu, Zn, α-tocopherol, and Fe-binding capacity. Whole blood samples were transferred
directly to microcentrifuge tubes containing an anticoagulant and frozen at -80°C until analysis
for glutathione peroxidase (GSH-Px) activity.
Liver samples were taken by biopsy 2 h before treatments were administered and then
randomly among the 6 cows between 14 to 16 h after treatment administration and 36 h posttreatment administration. Liver biopsies were performed as outlined by Bradford and Allen
(2005). At collection, cores were examined visually for contamination by adipocytes or excess

39

blood, and contaminated cores were discarded. Samples were stored in microcentrifuge vials at 10°C until analysis for Fe, Cu and Zn concentrations.

Common Feeding, Water and Milk Data Collection and Analysis.
A basal diet formulated to meet or exceed NRC (2001) recommendations for all nutrients
(including Fe) was fed ad libitum in the initial adjustment phase beginning 2 wk before period 1
through the conclusion of period 3 (Table 2.4). Cows were fed once daily in their individual tiestalls, and milked twice daily. Feed intakes, water intakes and milk yield and composition
analysis are in Table 2.1.
Feed and water intakes (in-line water flow meters) were measured daily throughout the
experiment (preliminary period through the end of period 3). Individual feed ingredients were
sampled the week before each collection period began, and again between experimental periods.
Each individual feed sample was weighed and dried in a forced-air oven at 55°C for 72 h until
completely dry to determine DM percentage. Dried samples were ground through a 2 mm screen
in a Wiley mill (Arthur H. Thomas, Philadelphia, PA). Similar feed ingredient samples were
pooled by experiment and sent for analysis by Dairy One Laboratories (Ithaca, NY). Milk
samples were taken weekly from one AM and PM milking and sent for composition analysis at
Michigan DHIA Universal Laboratory Services (Lansing, MI). Drinking water was sampled
daily from an in-line valve from the beginning of the preliminary period through the end of
period 3. One sample each from the beginning and end of each experiment were analyzed by
each of two separate methods for Fe. Raw water samples were sent to Midwest Laboratories

40

(Omaha, NE) for direct metals water analysis (Livestock water analysis, EPA 200.7 method
(1994)). A second sample was acidified (pH < 2.0) with 1 mL nitric acid, and sent to the same
laboratory for acid digestion and metals analyses (Total recoverable metals analysis using EPA
200.7 method (1994)). Water analyses are presented in Table 2.5. Internal water reference
samples containing either 0, 1, 4 and 8 mg of Fe/L from ferrous lactate were placed as unknowns
within the set of samples in triplicate for analysis. The intra-assay %CV was 4.0%.

Laboratory Analyses
Internal reference sample. Blood was taken from cows not in the experiments and serum
was harvested using the same procedure as that from experiment cows. This pool of serum and
whole blood was used as an internal reference sample to assess the variability within and among
all assays. Internal reference samples were placed in assay sets every 10 samples to comprise
approximately 10% of total samples analyzed. Once results were obtained for each analyte, the
percent coefficient of variation (% CV) was calculated, and if the % CV was greater than 10, the
samples, including additional internal reference samples, were re-analyzed.
Serum non-heme Fe and Fe binding capacity analyses. Analyses of serum Fe,
unsaturated Fe binding capacity (UIBC), total Fe-binding capacity (TIBC) and percent Fe
saturation were completed at the College of Veterinary Medicine Clinical Pathology Laboratory
at Michigan State University. The analyses were completed on an Olympus AU640e chemistry
immuno-analyzer (Olympus America Inc., Center Valley PA) utilizing Olympus reagents. The
serum samples were incubated in an acidic medium and Fe dissociated into Fe

41

3+

and apo-

transferrin molecules. Hydrochloric acid and sodium ascorbate were added to reduce Fe

3+

to

2+

Fe . Ferrous ions reacted with chromogen TPTZ ([2,4,6-Tri-(2-pyridyl)-5-triazine] to form a
complex measured at a wavelength of 600/800 nm.
Serum UIBC was measured by adding a known amount of Fe

3+

to a known sample

amount to saturate available transferrin binding sites. The remaining unbound Fe reacts with
nitroso-PSAP [2-nitroso-5-(N-propyl-N-sulfopropylamino)phenol] to form a complex measured
colorimetrically. The TIBC is calculated as: Fe (µg/dL) + UIBC (µg/dL). Percent Fe saturation is
calculated as: [Fe (µg/dl) / TIBC (µg/dL)] x 100. The intra- and inter-assay variation of the
internal reference samples were: for Fe: 2.19% and 1.88%; and for UIBC: 6.11% and 5.5%,
respectively.
Serum trace element analyses. Two hundred µL of each serum sample diluted with 5 mL
of a solution containing 0.5% EDTA and Triton X-100, 1% ammonia hydroxide, 2% propanol
and 20 ppb of scandium, rhodium, indium and bismuth as internal standards. An Agilent 7500ce
(Agilent Technologies, Santa Clara, CA) Inductively Coupled Plasma-Mass Spectrometer (ICPMS) was used for the analysis. Each element was calibrated using a 4-point linear curve of the
analyte: internal standard response ratio. The intra- and inter-assay variations of the internal
reference samples were: for Cu: 3.43% and 5.38% and for Zn: 3.94% and 5.45% respectively.
Serum α-tocopherol analysis. Fat-soluble vitamins were extracted from serum before
being analyzed by liquid chromatography. Serum (0.5 mL) was diluted with deionized water
containing 0.9% saline. Twenty µl of apocarotenal (as an internal standard), 1 ml of ethanol
containing butylated hydroxytolulene and 1 ml of hexane were added to the diluted serum. Caps

42

were replaced and the samples were mixed with a vortex mixer for 10 min. Samples were then
placed in a centrifuge and spun (1,000 g) for 10 min. After centrifugation, a positive
displacement pipette was used to remove 0.5 ml of the top layer of hexane, which was then
placed in a clean test tube and evaporated to dryness in a Rapidvap vacuum evaporation system
(Labconco, Kansas City, MO), set at 95 rpm, and 35°C. Once dry, the residue was suspended
with 0.5 ml of mobile phase, consisting of 70% acetyl nitrile, 10% methanol and 20% methylene
chloride. Samples were transferred to vials for analysis. Samples were analyzed by high
performance liquid chromatography (HPLC) at a wavelength of 292 nm. The sample (50 µL)
was injected in to a column (4.6 x 77 mm, Waters Corporation, Milford, MA) at a flow rate of
1.2 mL/min into a photodiode array detector (Waters 996 Detector; Waters Corporation, Milford,
MA). The absorbance was measured and area under the curve quantified. Quantification was by
internal standard ratio and a multipoint calibration curve. The intra- and inter-assay variations of
the internal reference samples were 6.21% and 6.77% respectively.
Whole blood GSH-Px activity analysis. Samples were thawed in a refrigerator for 70
min and analyzed in triplicate. Sixty-six µl of double Drabkins solution (400 mg potassium
cyanide, 79.2 mg potassium ferricyanide, and 200 mL distilled, deionized water), 714 µL of
distilled water, and 20 µL of sample were added to each culture tube in an ice bath, and mixed
using a vortex mixer. First, 270 µl of reaction buffer was added to each well of a 96-well plate,
consisting of 0.2M K2PO4 buffer, sodium azide solution, deionized water, 1.11 mg/dL NADPH,
and glutathione reductase enzyme. Fifteen µL of 12 mg/mL reduced glutathione and 12 µL of
diluted sample were added to each well. The well plate was incubated for 5 min at 25°C. After

43

incubation 3 µL of 12 mM H2O2 was added to each well, and the plate was read immediately
using a Spectramax plate reader (Molecular Devices, Sunnyvale, CA) at 340 mn, every 9 s for 5
min. Hemoglobin was measured by adding 5 mL of cyanmethemoglobin to 20 µL of whole
blood. The sample optical density was measured at 540 nm on a hemophotometer (Fisher
Scientific Co, Pittsburgh, PA). Results were reported as g/100mL of blood. The units of GSHPx activity were expressed as IU/g of hemoglobin. One IU is defined as the amount of enzyme
+

that will cause the oxidation of 1.0 nmol of NADPH to NADP per minute at 25°C. The interassay variation for the internal reference samples was 7.3%.
Liver tissue mineral element analysis. All liver tissue samples were prepared and
analyzed at the same time. Samples were thawed and placed in a 24°C oven to dry overnight.
Samples were removed and weighed. Three hundred µl of concentrated nitric acid were added to
each sample and left overnight to digest in an oven at 32-37°C. After acid-digestion samples
were diluted with deionized water to approximately 100 times their dry mass. After dilution, 200
µl of each sample were pipetted and diluted with 5 mL of a solution containing 0.5% EDTA and
Triton X-100, 1% ammonia hydroxide, 2% propanol and 20 ppb of scandium, rhodium, indium
and bismuth as internal standards. An Agilent 7500ce (Agilent Technologies, Santa Clara, CA)
Inductively Coupled Plasma-Mass Spectrometer (ICP-MS) was used for the analysis.

Statistical Analyses
Data were analyzed by method of least-squares ANOVA using the Mixed Models
procedure of SAS (Version 9.2, SAS Institute Inc.). Data of hourly measurements were analyzed

44

using the REPEATED statement, and the covariance structure that resulted in the lowest
Bayesian information criterion (BIC) was used. Fixed effects in the model were treatment, hour
and treatment by hour interaction. Random effects were cow and period. If the interaction was
non-significant (P > 0.15), it was removed from the model. Additionally, if hour was nonsignificant (P > 0.10), it also was removed from the model. Mean of the pre-dosing samples from
each cow was used as a covariate for each dependent variable. If the covariate term was not
significant, it was removed. Residual distributions were tested for normality using the ShapiroWilk test, and outliers were identified using Cook’s D statistic. Differences between treatments
were identified using the PDIFF option with Tukey’s adjustment. Least-squares means and SEM
are presented. Significance for main effects was declared at P < 0.05, and a trend toward
significant difference at P < 0.10. Significance for interaction effects was declared at P < 0.10,
and a trend toward significance at P < 0.15. Pearson correlations were calculated for the
relationships between independent response variables using the Correlation procedure.

45

RESULTS
Water Analyses
Average water sample Fe concentrations were 0.42 and 0.46 mg/L from direct Fe
analysis, and 1.59 and 2.36 mg/L from total recoverable Fe analysis for Experiments 1 and 2,
respectively. The Fe concentrations from the direct Fe analysis were approximately 25% of the
total recoverable Fe concentrations. The direct Fe analysis is that used in the field to evaluate
water quality in dairy farms (Beede, personal communication, Michigan State University, 2011).

Treatment by Hour Interactions
There were no treatment by hour interactions for any response variable. There was a
tendency for a treatment by hour interaction in hepatic Cu concentration (P = 0.10) in
Experiment 1. The 0.75Fe treatment tended to cause a greater decrease in hepatic Cu
concentration than other treatments (Figure 2.1).

Experiment 1
Treatment effects. Our objective was to characterize the effects of varying amounts of
ferrous lactate abomasally infused on Fe and antioxidant status. Lactating cows were
administered 0, 0.75, or 1.5 mg of Fe from ferrous lactate per kg of BW, and Fe status responses
were evaluated for 12 h. We found no effects of treatment on serum Fe, UIBC, Fe saturation, Zn
or Cu, whole blood GSH-Px activity, or hepatic Fe and Zn concentrations (Table 2.6). Serum
TIBC tended to be greater for cows on 0.75 Fe than cows on 0Fe (P = 0.06). Serum α-tocopherol

46

was greater in response to 1.5Fe than 0Fe (P = 0.006), and hepatic Cu concentrations were lower
in cows administered 0.75Fe than cows administered 0Fe (P = 0.02).
Hourly effects. When data were pooled across treatments there were some effects of hour
on response variables. Both serum Fe (P = 0.04; Figure 2.2) and Fe saturation (P = 0.01; Figure
2.3) decreased over time. Serum Zn concentrations decreased over the experimental period (P <
0.0001; Figure 2.4). Hepatic Cu concentrations decreased over time (P= 0.02; Figure 2.5),
whereas hepatic Zn tended to decrease from 0 h to 36 h (P = 0.09; Figure 2.6). There were no
hourly effects on serum TIBC, UIBC, α-tocopherol and Cu, whole blood GSH-Px activity, and
hepatic Fe concentrations.
Correlations. The correlations between various independent serum and liver response
variables were investigated (Tables 2.7, 2.8). There was a negative correlation between serum Fe
and serum UIBC (r = -0.46; P < 0.05).

Experiment 2
Treatment effects. The objective was to evaluate the effects that Fe

2+

3+

or Fe

have on Fe

and antioxidant status of lactating dairy cows. Iron was supplied as 0 mg of Fe or 1.5 mg of
either ferrous or ferric sulfate per kg BW. Overall there were no effects of treatment on serum
TIBC, UIBC, Fe saturation, α-tocopherol, Zn; whole blood GSH-Px; or hepatic Fe, Cu or Zn
concentrations (Table 2.9). Serum Cu concentrations in response to ferrous sulfate were greater
than serum Cu concentrations for cows on either Control or ferric sulfate treatments (P =

47

0.0002). Serum Fe concentrations tended to be lower for cows on the ferric sulfate treatment
than Control cows (P < 0.10).
Hourly effects. Hourly effects pooled across treatment were found for serum Fe, UIBC,
Fe saturation, α-tocopherol, Zn, and whole blood GSH-Px. Serum Fe decreased over the
experimental period (P = 0.01; Figure 2.7) and UIBC (P = 0.001; Figure 2.8) and Fe saturation
(P < 0.0001; Figure 2.9) also decreased. Serum Zn decreased slightly, but consistently
throughout the experimental period (P <0.0001; Figure 2.10). There was an hour effect on αtocopherol (P = 0.001; Figure 2.11) though no consistent trend to increase or decrease was
identified. Whole blood GSH-Px tended to decrease from 0 h to 6 h, but then increase slightly
by h 12 (P = 0.07; Figure 2.12).
Correlations. The correlation between various independent serum and liver response
variables was investigated (Tables 2.10, 2.11). There was a correlation between serum Fe and
serum UIBC (r = -0.54; P < 0.05) and hepatic Cu and Zn (r = 0.6535; P < 0.05).

48

DISCUSSION
In our study, 0.75 mg Fe/kg BW and all 1.5 mg Fe/kg BW treatments were formulated to
contain approximately the total amount of Fe that a 600-kg lactating cow drinking 100 L of
water/d would consume if drinking water with total Fe concentrations of 4.5 and 9 mg/L,
respectively. However, after receipt of water analyses we realized that these values are not
necessarily accurate. The laboratory utilizes the EPA 200.7 method for direct metals for
Livestock water analysis. However, there is an additional method that can quantify total
recoverable Fe through the use of acid digestion. After receiving the two types of analyses, we
noticed that the Fe values reported for the direct metals analysis were approximately 25% of the
values for the total recoverable metals analysis. Based on this information, we suspect that the
concentrations of Fe dosed may be more similar to the total Fe intake of an average cow drinking
100 L of water daily with Fe concentrations of 0, 1, or 2.25 mg/L respectively. This could be
reason for anecdotal evidence that high concentrations of Fe in drinking water cause problems in
cow health and productivity. Drinking water containing 2 mg Fe/L may actually contain 8 mg
Fe/L.

Experiment 1
Serum Fe concentrations. The serum Fe concentrations of cows dosed with ascending
amounts of ferrous lactate were not affected. This was unexpected as humans orally administered
similar amounts of Fe

2+

as ferrous sulfate (approximately 1.4 mg Fe/kg BW) in solution had

increased serum Fe concentrations (approximately 100 µg/dL) within 2 h (Walker et al., 1989).

49

However, Ammerman et al. (1967) orally dosed sheep of adequate Fe status with 150 µC Fe

59

from either ferrous sulfate or carbonate, providing 1.8 mg total Fe/kg BW, or ferric chloride or
oxide, providing 0.8 mg total Fe/kg BW. They reported that the radio-labeled Fe could be
detected in the serum and was maximal at 12 h post-dosing, independent of source, though no
more than 0.25% of labeled Fe was detected in the serum from any source. It is likely that the
systems involved in Fe regulation can easily handle small increases in Fe intake. Other research
showed that serum Fe did not increase in response to a meal in humans (Sinnah et al., 1969).
Hansen et al. (2010) reported that although calves supplemented with 750 mg Fe/kg
dietary DM of ferrous sulfate (approximately 22 mg Fe/kg BW) did not have increased
concentrations of Fe in serum, Fe addition negatively affected ferroportin expression, and
increased hepcidin expression in intestinal tissue. The results of our experiment suggest that the
concentrations of Fe contained in poor quality water apparently are not great enough to
overwhelm the regulation of Fe absorption and significantly affect short-term Fe status.
It is possible that pH of the treatment solutions had an effect on Fe absorption. The 1.5Fe
treatment had a greater pH, as it did not contain supplemental acid (Table 2.2). Between the pH
range of 4 to 8, as pH of an aqueous solution decreases, Fe

2+

2+

and the ratio of Fe /Fe

solution increases (Morgan and Lahav, 2007). At a lower pH, the concentration of Fe
greater, and Fe

2+

3+

2+

in the
is

3+

is more available for absorption than Fe . Solomans and Jacob (1981)

administered 1 g ascorbic acid with ferric chloride to increase Fe absorption in humans. In rats
administered cimetidine, a compound that inhibits gastric acid secretion, absorption of Fe was
decreased (Kim et al., 1993) indicating that an acidic environment is essential for Fe absorption.

50

In our experiment the greater pH of the 1.5Fe treatment could explain the lack of difference
between serum Fe concentrations of cows given 0.75Fe and 1.5Fe treatment even though 1.5Fe
contained twice as much Fe.
Serum TIBC, UIBC and percent Fe saturation. Serum TIBC is an indirect measurement
of total Fe-binding proteins present in the serum (mainly transferrin). Transferrin is the binding
protein for Fe transport in serum. Serum UIBC is an indirect measurement of the Fe binding
proteins present in the serum that are not bound to Fe. In general, there is an inverse relationship
between serum Fe and serum UIBC (Furugouri, 1971). This was true in Experiment 1. There was
a negative correlation between serum Fe and UIBC (r = -0.46; Table 2.7). Cows administered
0.75Fe tended to have greater serum TIBC than 0Fe-cows (Table 2.6). Serum Fe was
numerically greater, and UIBC numerically lower for 0.75Fe cows, suggesting that this treatment
had the greatest overall impact. This explains the treatment difference, as TIBC is reflective of
both serum Fe and UIBC.
Serum and hepatic Cu concentrations. Copper is transported into the enterocyte
through two main transporters, DMT-1 and copper transporter 1 (Ctr1) (Sharp, 2004). DMT-1
2+

has a high affinity for Fe , and when excess Fe is present it can prevent the binding of Cu to
DMT-1, reducing Cu absorption (Sharp, 2004). Decreased absorption could cause depletion in
liver Cu reserves, as Cu stores are mobilized to maintain Cu homeostasis. In Experiment 1 it was
expected that 1.5Fe would decrease serum and liver Cu most, because it delivered the most Fe;
however, this was not the case. There was a tendency for a treatment by hour interaction in
hepatic Cu concentration, as 0.75Fe decreased hepatic Cu concentration over time to a greater

51

extent than other treatments (Figure 2.1), and cows administered 0.75Fe had lower liver Cu
concentrations when data were pooled across hour (Table 2.6). According to Standish et al.
(1969) when cattle were administered high concentrations of dietary Fe as ferrous sulfate, there
was a decrease in hepatic Cu concentration. In Experiment 1, 0.75Fe caused a greater decrease in
hepatic Cu concentration than other treatments (Table 2.6).
Serum and hepatic Zn concentrations. There was no effect of treatment on serum or
liver Zn concentrations. There is some evidence for an interaction between Fe and Zn in
ruminants; however this relationship is less well-formulated and consistent than the relationship
between Fe and Cu (Campbell et al., 1974, Standish et al., 1969, Standish and Ammerman,
1971). It is likely that the interaction only manifests when animals have been exposed to high
concentrations of Fe for a long period of time. Steers fed high dietary concentrations of ferrous
sulfate (0, 400, 1,600 or 3,200 mg Fe/kg), had decreased liver Cu and Zn concentrations as
dietary Fe concentration increased (Standish et al., 1969). Steers were exposed to high
concentrations for 49 d before sampling. We suspect that over a longer period of time than
Experiment 1, exposure to Fe could impact Zn status.
Antioxidant status and oxidative stress. In vitro, during Fe-induced oxidative stress, αtocopherol concentration decreased rapidly (Yamamoto and Niki, 1988). Somewhat in contrast,
in our study serum α-tocopherol concentrations of cows on 0Fe were lower than that of cows on
1.5Fe. Additionally we found no change in GSH-Px activity in response to treatment. In the
body, GSH-Px and vitamin E work in concert to protect tissues from oxidative damage induced
by Fe (Milchak and Bricker, 2002). Also, induced α-tocopherol deficiency decreased glutathione

52

(an antioxidant) and increased hydroperoxides in rats (Nickander et al., 1994). Both the lack of
response of GSH-Px activity and increase in serum α-tocopherol in Experiment 1 suggest that
supplemental Fe did not induce oxidative stress. Similarly, in a study examining the relationship
of supplemental Fe and antioxidant status, 1,000 mg/kg of supplemental Fe from ferrous sulfate
added to swine starter diets had no effect on either GSH-Px activity or serum α-tocopherol
concentrations (Dove and Ewan, 1990). Hansen et al. (2010) reported no change in GSH-Px
activity in response to high dietary Fe from ferrous sulfate (750 mg/kg) in liver or heart tissue.
Our results suggest that the amounts of Fe administered in Experiment 1 were not enough to
impact antioxidant status, or cause oxidative stress short-term.

Experiment 2
Serum Fe concentrations. We expected serum Fe concentrations of cows on both ferrous
and ferric sulfate treatments to be greater than those on the Control treatment, and serum Fe
concentrations of cows given ferrous sulfate to be greatest. In contrast, there was a tendency for
serum Fe to be lower in cows infused with ferric sulfate compared with Control, and cows given
ferrous sulfate were not different from cows given Control or ferric sulfate treatments (Table
2.9). As mentioned previously, Fe absorption is tightly regulated, and it is not known if Fe
3+

Fe

3+

affect transporters involved in Fe metabolism differently. It is possible that Fe
2+

2+

has a
3+

similar effect on the presence of DMT-1 in the apical membrane as Fe . However as Fe

must

be reduced in the intestinal lumen before absorption, if there are fewer transporters available,

53

or

2+

absorption could be decreased more than in response to Fe . Decreased Fe absorption, and Fe
export from the enterocyte, would result in lower serum Fe concentrations short-term.
Serum TIBC, UIBC and percent Fe saturation. There was no effect of treatment on
concentrations of serum TIBC, UIBC or percent Fe saturation (Table 2.9). In contrast, calves fed
1,000 mg Fe/kg dietary DM supplemental ferrous carbonate had decreased serum UIBC and
TIBC after 1 wk, but recovered to pre-feeding values after 2 wk (Ho et al., 1984). We expected
ferrous sulfate to result in lower UIBC, and increase TIBC and Fe saturation percentage, but 1.5
mg Fe/kg BW from either Fe source did not impact measurements of Fe status in Experiment 2.
Serum and hepatic Cu concentrations. There was an effect of treatment on serum Cu
concentrations. Both cows on the Control and ferric sulfate treatments had serum Cu
concentrations that were greater than those on ferrous sulfate, and there were no differences in
hepatic Cu concentrations (Table 2.9). Standish and Ammerman (1971) reported that wethers
administered 1,600 mg Fe/kg dietary DM of ferrous sulfate had decreased serum concentrations
of Cu, whereas the same dietary concentration of ferric citrate had no effect, similar to the results
in our experiment. Campbell et al. (1974) found that cows intraruminally dosed weekly with 210
mg/kg BW of ferric hydroxide slurry had markedly decreased hepatic Cu stores, and serum Cu
2+

concentrations, but not liver Zn. The authors suggested that though Fe
than ferric iron, Fe

3+

is more bioavailable

is more effective at rapidly depleting hepatic Cu. This was not the case in

our experiments as hepatic Cu concentrations of cows given ferric sulfate were not different from
cows administered the Control treatment. It is likely that the difference between the current
experiment and that of Campbell et al (1974) was due to different amounts of Fe administered.

54

Serum and hepatic Zn concentrations. Whereas there is some evidence for an
interaction between Fe and Zn in absorption, the mechanisms are not known, and as mentioned
previously, the interaction of Fe and Zn has not been demonstrated as consistently as that of Fe
and Cu (Campbell et al., 1974, Standish et al., 1969, Standish and Ammerman, 1971). In
Experiment 2 there was no treatment-related difference in either serum or hepatic Zn
concentrations. Solomans and Jacob (1981) concluded that while both Fe

2+

3+

and Fe

(50 mg as

ferrous sulfate or ferric chloride) administered to humans decreased absorption of a 25 mg dose
2+

of Zn, Fe

3+

inhibited Zn uptake to a greater extent than Fe . Our experimental results are not

consistent with this observation, as we found no difference between Control, ferrous or ferric
sulfate on Zn measurements. This is in agreement with Experiment 1, as amount of Fe

2+

dosed

also did not impact Zn status. Results from both Experiments 1 and 2 indicate Zn was not
impacted by the amounts and valences of Fe administered in our studies.
In Experiment 2 we found a positive correlation between hepatic Cu and Zn
concentrations (Table 2.11). This effect was likely heavily influenced by 2 individual cows that
had markedly greater concentrations of Cu and Zn than other cows. A similar relationship was
not found in Experiment 1.
Antioxidant status and oxidative stress. There was no effect of ferrous or ferric sulfate
treatments on serum α-tocopherol concentrations or whole blood GSH-Px. Lack of treatment
differences suggests that there may be no difference between Fe supplied through either feed or
water short-term, on antioxidant status and oxidative stress. Braughler et al. (1986) found that
3+

Fe

alone did not induce lipid peroxidation of rat brain tissue in vitro. Although we expected

55

that ferrous sulfate would have a greater impact on oxidative stress than ferric sulfate, this was
not the case. At the concentrations dosed, neither Fe

2+

3+

nor Fe

causes oxidative stress short-

term.
We found a positive correlation between serum α-tocopherol concentration and whole
blood GSH-Px (Table 2.10). This suggests a potential relationship between the concentration of
α-tocopherol in the serum and GSH-Px; when α-tocopherol concentration was greater, enzyme
activity also was greater. This pattern is contrary to the results reported by Chow et al. (1973)
that rats fed deficient dietary concentrations of α-tocopherol had greater GSH-Px activity.
However, this pattern was not found in Experiment 1, indicating that additional research should
be completed in this area.

Hourly Effects Pooled Across Treatments for Experiments 1 and 2
In our experiments, because we sampled blood hourly for 12 h, we were able to
characterize changes in response variables over time; by treatments and pooled across
treatments. In both Experiments 1 and 2, there were no overall effects of hour pooled across
treatments for serum α-tocopherol, whole blood GSH-Px, and or Fe concentrations. In
experiment 1 we found no overall effect of hour pooled across treatments on serum TIBC, UIBC,
α-tocopherol, and Cu concentrations, whole blood GSH-Px activity or hepatic Fe concentrations.
In Experiment 2 there was no overall hourly effect on serum TIBC and Cu concentrations, and
hepatic Fe, Cu and Zn concentrations.

56

Serum Fe and UIBC concentrations. There is very little research on diurnal variation in
serum and plasma Fe concentrations in cattle. Mansion et al. (1981) evaluated the effect of time
of day on variation in serum Fe in Friesian cows and heifers, and found that there was no
significant pattern throughout the day. These researchers concluded that the main source of
variation was the animal.
This is contrary to our findings in Experiments 1 and 2. In Experiment 1 serum Fe
decreased from h 2 (2000 h) through the rest of the sampling period (0600 h), indicating that
daily serum Fe was greatest before 2000 h (Figure 2.2). This was similar to results in Experiment
2 in which serum Fe concentrations decreased from h 4 (2200 h) through h 11 (0500 h) (Figure
2.7). Our findings are similar to those in humans completed by Wiltink et al. (1973). Those
researchers evaluated changes in plasma Fe over a 24 h period and determined that
concentrations were greatest from 1200 to 2000 h and lowest between 2400 and 0800 h. Diurnal
variations in serum Fe concentrations in humans are not affected by food intake (Sinnah at al,
1969). Stengle and Schade (1957) found no significant variation in TIBC in humans, which is in
agreement with both of our experiments. In Experiment 2, serum UIBC increased from hour 1
(1900 h) through hour 11 (0500 h) (Figure 2.8). This effect is likely related to the serum Fe
decrease over time because those responses are correlated as shown in Experiments 1 and 2.
Serum percent Fe saturation. Serum percent Fe saturation is an indirect measurement of
Fe saturation of transferrin and is calculated as percent saturation = [serum Fe/TIBC] x 100. It
would be expected that as serum Fe concentrations changed, there would be a concurrent change
in serum percent saturation. We found that serum Fe (Figures 2.2, 2.7) and percent Fe saturation

57

(Figures 2.3, 2.9) decreased over time in both experiments. Overall, our data suggest there may
be a pattern in daily serum Fe concentration, serum UIBC, and percent saturation in lactating
dairy cows.
Serum Cu concentrations. There was no hourly effect on serum Cu concentrations in
either experiment. Mansion et al. (1981) found that serum Cu increased linearly from 0700 to
1600 h in dairy cows. Diurnal variation of serum Cu in cattle should be investigated further.
Serum Zn concentrations. Although no information exists about diurnal variation in
cattle, studies completed in humans suggest that plasma Zn concentrations decrease from 0800 to
1900 h and increase overnight (McMillan and Rowe, 1982; Markowitz et al., 1985). Also daily
variation may be linked to food intake patterns, but not Zn content of the diet (Hashim et al.,
1996). This is different than our findings. There was a clear pattern in serum Zn concentrations
in both experiments; serum concentrations decreased overnight (from 1900 to 0600 h) (Figures
2.4, 2.10). It is possible that serum concentrations peaked shortly after feeding (1100 h) and
decreased slowly throughout the day. A thorough evaluation of diurnal variation in cattle is
needed.
Serum α-tocopherol concentrations. There is limited research on the diurnal variation of
fat-soluble vitamins, but Nierenberg and Stukel (1987) found that hourly variation of plasma
tocopherols was minimal, and non-significant in humans. No previous research evaluating the
variation of plasma or serum fat-soluble vitamin concentrations in cattle or other ruminants was
found. There was no hourly effect on α-tocopherol in Experiment 1. There was an effect in

58

Experiment 2 (Figure 2.11), however the effect was influenced mainly through variability from
hour to hour rather than being a consistent effect over time.
Whole blood GSH-Px activity. There was no effect of hour on whole blood GSH-Px
activity in Experiment 1. In Experiment 2 there was a tendency for activity to decrease from h 0
(1800 h) to h 6 (0000 h) and then increase slightly at the end of the experimental period (Figure
2.12). In human subjects limited research suggests that glutathione concentrations are subject to
diurnal variation; concentrations are maximal around 0230 h, and minimal around 1330 h.
Information about daily variation in GSH-Px activity was not found. Plasma malonyldialdehyde
(MDA) concentrations generally are found to be greatest from 1200 to 1600 h and lowest at 0400
h in healthy subjects (Akbulut et al., 2003), suggesting that lipid peroxidation follows a diurnal
pattern. This is partial evidence for a pattern in GSH-Px activity, however verification of this
potential pattern requires additional research.
Hepatic trace mineral concentrations. There is no information in the literature about
variation in hepatic mineral concentrations during such a short time period, as no other
researchers have taken liver tissue samples as frequently as in our study. In Experiment 1, there
was a decrease in hepatic Cu (Figure 2.5) and a tendency for a decrease in hepatic Zn (Figure
2.6) from h 18 (1200 h) to h 36 (0600 h), whereas there was no change in hepatic Fe
concentration. These patterns were not observed in Experiment 2. Additional work should be
completed in this area in order to explore a potential pattern in variation.

59

CONCLUSIONS
2+

In conclusion, there was little difference between 0, 0.75 and 1.5 mg Fe /kg BW, and
2+

1.5 mg Fe

3+

or Fe /kg BW of on measurements of Fe and antioxidant status, or oxidative stress.

There was some impact on serum and hepatic Cu concentrations, though the effects were
variable and warrant further investigation.
Overall, the amounts of Fe administered abomasally in these experiments apparently are
not great enough to affect short-term Fe status. However, we suggest that chronic exposure to
greater amounts of Fe in drinking water, or exposure to greater concentrations of Fe, could cause
antioxidant depletion and oxidative stress indicative of situations described in the field. This is
supported by the information discovered in our work that the actual total recoverable
concentrations of Fe in drinking water are about 4 times greater than conventionally believed
based on the routinely used Livestock water analysis. These much greater concentrations could
lead to decreased water intake, feed intake and milk production. Alternatively, decreased
drinking water palatability from Fe contamination, or a combination of both of these effects
could cause a decrease in cow production and health.

60

APPENDIX B
FIGURES AND TABLES REFERENCED IN CHAPTER 2

61

TABLES
Table 2.1. Pre-experimental dry matter intake, milk yield, water intake and body weight of cows
in Experiments 1 and 2
Experiment
1

2

Mean
DMI, kg

±SD

1

Mean

±SD

23

1.9

25

1.6

Water intake, L

79.6

30.2

102.5

7.6

Body weight, kg

768

82.5

658

51.2

Milk yield, kg

36.2

13.7

36.9

3.5

Milk composition, %
Fat

3.10

0.56

3.68

0.48

True protein

2.90

0.11

2.99

0.12

Lactose

4.51

0.36

4.92

0.09

5.37

0.42

5.85

0.10

SNF

2

3

SCC , 1000 ml
4

MUN , mg/dl
1
2
3
4

186

112.0

17.1

1.56

SD = standard deviation.
SNF = solids not fat.
SCC = somatic cell count.
MUN = milk urea nitrogen.

62

258
14.7

357.8
1.71

Table 2.2. Characteristics of treatment infusion solutions in Experiment 1
Treatments
0Fe
Solution characteristics

0.75Fe

Mean

±SD

Total Volume, L

1.00

pH
Lactate, g
Fe, g
1

1

2

1.5Fe

Mean

±SD

Mean

±SD

--

1.00

--

1.00

--

2.56

0.09

3.73

0.06

5.31

0.03

3.70

0.39

3.73

0.40

3.71

0.41

--

--

0.58

0.06

1.15

0.13

Treatments: 0Fe = 0 mg Fe/kg BW; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate

[Fe(C3H5O3)2]; 1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

SD = standard deviation.

Table 2.3. Characteristics of treatment infusion solutions in Experiment 2
Treatments
Control
Solution characteristics

Mean

±SD

1

ferrous sulfate
2

ferric sulfate

Mean

±SD

Mean

±SD

Total volume, L

1.50

0.00

1.50

0.00

1.50

0.00

pH

1.63

0.06

2.12

0.08

2.48

0.11

Sulfate, g

2.54

0.19

2.54

0.20

2.54

0.19

Fe, g

0.00

0.00

0.99

0.08

0.98

0.08

1

Treatments: Control = no Fe; ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate [FeSO4];

ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

SD = standard deviation.

63

Table 2.4. Ingredient and nutrient composition of diets fed in Experiments 1 and 2
Experiment
1
2
Ingredient
--------------------Percent of dietary DM-------------------Chopped alfalfa hay
4.86
4.86
Corn silage
32.24
32.96
Whole cottonseed
5.95
5.86
Grass hay
2.08
2.11
Ground corn
14.42
14.29
Alfalfa haylage
9.14
8.68
High moisture corn
6.29
6.35
Soybean meal, 48% CP
14.04
13.74
Soyhulls
7.77
7.97
1
Mineral-vitamin mix
3.22
3.18
Nutrient Composition
DM %
52.9
53.80
--------------------Percent of dietary DM-------------------Neutral detergent fiber
31.4
32.0
Acid detergent fiber
21.3
18.3
Crude protein
17.51
17.51
Ca
0.88
1.01
P
0.36
0.33
K
1.23
1.17
Na
0.30
0.31
S
0.22
0.22
Mg
0.22
0.24
--------------------------mg/kg----------------------------Co
0.12
0.12
Cu
8.5
8.9
Fe
185
240
Mo
1.9
1.9
Mn
37
36
Zn
30
38
1

Mineral-vitamin mix contained 47.5% limestone, 22.5% sodium bicarbonate, 10.1% urea

(45%N), 8.2% magnesium sulfate, 7.5% sodium chloride, 1.6% biotin premix (1.4 g/kg), 1.3%
trace mineral premix, 30 KIU/kg vitamin A, 8 KIU/kg vitamin D, 56 KIU/kg vitamin E and 0.3%
selenium yeast, dry basis.
64

Table 2.5. Water quality constituent analyses of drinking water in Experiments 1 and 2
Experiment
1

2

Quality Constituent
TDS

Caution Level

--------------------mg/L--------------------

2

392

454

1,000

Ca

94.7

104

150

Cl

9

27.5

500

ND

0.3

0.42

0.46

0.3

1.59

2.36

-

33

36

Mn

0.05

0.15

NO3-N

ND

ND

25

Na

5.2

14.8

150

SO4

65

101

300

Zn

1.33

1.02

-

0.603

0.698

1.5

7.9

8.1

6.5-9

Cu

ND

Fe (direct)

4

Fe (total recoverable)

5

Mg
4

4

Conductivity, mmhos/cm
pH
1
2
3
4

3

80
-

Caution level from Midwest Laboratories (Omaha, NE).
TDS = total dissolved solids.
ND = not detected.
Direct metals analysis of raw (without acid digestion) water by Livestock water analysis using

EPA 200.7 method (1994) by Midwest Laboratories (Omaha, NE).
5

1

Total recoverable metals analysis of acidified water after acid digestion using EPA 200.7

method (1994) by Midwest Laboratories (Omaha, NE).

65

Table 2.6. Effects of abomasal dosing of different amounts of Fe

2+

from ferrous lactate on blood

and liver measurements in Experiment 1
Treatments

1
2

0Fe

0.75Fe

1.5Fe

SEM

P-value

178

192

183

15.7

0.71

Serum
Fe, µg/dL
3

TIBC , µg/dL

†

469

UIBC , µg/dL
5

Fe saturation , %

485

478

13.6

0.08

301

4

293

298

19.2

0.35

40

39

3.7

0.92

6.085

0.2170

0.01

38
b

ab

5.894

a

α-Tocopherol, µg/mL

5.694

Zn, µg/mL

0.993

0.987

0.947

0.0385

0.66

Cu, µg/mL

0.991

1.003

0.979

0.0305

0.49

2.46

2.54

2.49

0.051

0.47

287.9

281.0

273.7

33.88

0.69

a

b

19.63

0.03

14.01

0.64

Whole blood
6

GSH-Px , IU/g Hb

7

Liver, µg/g dry weight
Fe
Cu

703.3

639.2

Zn

134.3

133.9

a,b
1

ab

663.0

148.2

Means within a row with different superscripts differ (P < 0.05).

Treatments: 0Fe = 0 mg Fe/kg BW; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate

[Fe(C3H5O3)2]; 1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2
3
4
5
6
7
†

SEM = standard error of the mean.
TIBC = total Fe binding capacity = Fe + UIBC.
UIBC = unsaturated Fe binding capacity.
Fe Saturation = (Fe/TIBC) x 100.
GSH-Px = glutathione peroxidase activity.
Hb = hemoglobin.
Tended to be lower (P <0.10) than 0Fe.

66

Table 2.7. Pearson correlations coefficients of relationships between independent serum
response variables pooled across treatment and hour for Experiment 1
1

2

UIBC

Zn

GSH-Px

-0.46214*

0.38375

0.21908

0.28852

-0.19542

UIBC

--

-0.37463

0.20208

0.04449

0.29176

α-Tocopherol

--

--

0.22133

-0.26678

0.01971

Cu

--

--

--

-0.08332

-0.01447

Zn
2

Cu

Fe

1

α-Tocopherol

--

--

--

--

-0.1946

UIBC = unsaturated Fe binding capacity.
GSH-Px = glutathione peroxidase activity.

*P < 0.05

Table 2.8. Pearson correlations coefficients of relationships between independent liver response
variables pooled across treatment and hour for Experiment 1
Cu

Zn

Fe

0.55385

0.49768

Cu

--

0.53844

Zn

--

--

67

Table 2.9. Effects of abomasal dosing of ferrous sulfate or ferric sulfate on blood and liver
measurements in Experiment 2
Treatments

1

ferrous

ferric

Control

sulfate

sulfate

169

161

425

2

SEM

P-value

149

11.7

0.11

420

426

4.2

0.36

254

263

274

9.5

0.22

40

38

35

2.6

0.14

α-Tocopherol, µg/mL

5.189

5.412

4.829

0.2258

0.13

Zn, µg/mL

0.852

0.807

0.772

0.0468

0.20

a

b

a

Serum
Fe, µg/dL
3

TIBC , µg/dL
4

UIBC , µg/dL
5

Fe saturation , %

Cu, µg/mL

†

1.006

0.950

1.036

0.0352

0.0002

2.24

2.19

2.24

0.065

0.46

Fe

348.6

338.6

380.1

27.75

0.21

Cu

647.5

629.3

661.1

25.54

0.31

Zn

152.3

154.1

164.1

7.818

0.28

Whole blood
6

GSH-Px , IU/mg Hb

7

Liver, µg/g dry weight

a,b
1

Means within a row with different superscripts differ (P < 0.05).

Treatments: Control = no Fe; ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate [FeSO4];

ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2
3
4
5
6
7
†

SEM = standard error of the mean.
TIBC = total Fe binding capacity = Fe + UIBC.
UIBC = unsaturated Fe binding capacity.
Fe Saturation = (Fe/TIBC) x 100.
GSH-Px = glutathione peroxidase activity.
Hb = hemoglobin.
Tended to be lower (P < 0.10) than ferrous sulfate.

68

Table 2.10. Pearson correlations coefficients of relationships between independent serum
response variables pooled across treatment and hour in Experiment 2
1

2

UIBC

Zn

GSH-Px

-0.54061*

0.1578

0.02096

0.30743

-0.04863

UIBC

--

-0.38257

0.10746

-0.34677

0.01364

α-Tocopherol

--

--

-0.10519

-0.03308

0.49837*

Cu

--

--

--

-0.13694

-0.29371

Zn
2

Cu

Fe

1

α-Tocopherol

--

--

--

--

-0.27009

UIBC = unsaturated Fe binding capacity.
GSH-Px = glutathione peroxidase activity.

* P < 0.05.

Table 2.11. Pearson correlations coefficients of relationships between independent liver response
variables pooled across treatment and hour in Experiment 2
Cu

Zn

Fe

0.04967

0.02190

Cu

--

0.80838*

Zn

--

--

* P < 0.05.

69

FIGURES
1

2

Figure 2.1. Treatment by hour interaction response (SEM = 26.55, P = 0.10) of hepatic Cu
concentrations in Experiment 1
800
Hepatic Cu, ug/g dry tissue

750
700
650
600
550

0Fe

500

0.75Fe

450

1.5Fe

400
350
300
0

18

36

Hour
1

Treatments: 0Fe = 0 mg Fe/kg BW; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate

[Fe(C3H5O3)2]; 1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

SEM = standard error of the mean.

70

1

2

Figure 2.2. Treatment by hour (P = 0.63) and pooled hour (SEM = 17.55; P = 0.04) responses
of serum Fe concentration in Experiment 1
220

Serum Fe. µg/dL

200
180
160
140
120

0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

100
1
1

2

3

4

5

6
7
Hour

8

10

11

12

Treatments: 0Fe = 0 mg Fe/kg BW; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate

[Fe(C3H5O3)2]; 1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

9

SEM = standard error of the mean.

71

1

2

Figure 2.3. Treatment by hour (P = 0.82) and pooled hour (SEM = 3.94; P = 0.008) responses
3

of serum Fe saturation in Experiment 1
45

Serum Fe Saturation, %

40

35

30

25
0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

20
1
1

2

3

4

5

6
7
Hour

8

3

10

11

Treatments: 0Fe = 0 mg Fe/kg BW; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate

[Fe(C3H5O3)2]; 1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

9

SEM = standard error of the mean.
Fe Saturation = Fe/total Fe binding capacity.

72

12

1

2

Figure 2.4. Treatment by hour (P = 0.87) and pooled hour (SEM = 0.06; P < 0.0001)
responses of serum Zn concentration in Experiment 1
1.2

Serum Zn, µg/mL

1
0.8
0.6
0.4
0Fe

0.2

0.75Fe

1.5Fe

Hour effect pooled across treatments

0
1

2

3

4

5

6

7

8

9

10

11

Hour
1

Treatments: 0Fe = 0 mg Fe/kg BW; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate

[Fe(C3H5O3)2]; 1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

SEM = standard error of the mean.

73

12

1

2

Figure 2.5. Treatment by hour (P = 0.62) and pooled hour (SEM = 26.55; P = 0.02) responses
of hepatic Cu concentration in Experiment 1
800

Hepatic Cu, µg/g dry tissue

700
600
500
400
300

0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

200
0

18

36

Hour
1

Treatments: 0Fe = 0 mg Fe/kg BW; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate

[Fe(C3H5O3)2]; 1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

SEM = standard error of the mean.

74

1

2

Figure 2.6. Treatment by hour (P = 0.99) and pooled hour (SEM = 7.14; P = 0.08) responses
of hepatic Zn concentration in Experiment 1
180
Hepatic Zn, µg/g dry tissue

170
160
150
140
130
120
110
100

0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

90
80
0
1

18
Hour

Treatments: 0Fe = 0 mg Fe/kg BW; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate

[Fe(C3H5O3)2]; 1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

36

SEM = standard error of the mean.

75

1

2

Figure 2.7. Treatment by hour (P = 0.85) and pooled hour (SEM = 13.03; P = 0.01) responses
of serum Fe concentration in Experiment 2
220

Serum Fe. µg/dL

200
180
160
140
120

Control
Ferric sulfate

Ferrous sulfate
Hour effect pooled across treatments

100
1

2

3

4

5

6

7

8

9

10

11

12

Hour
1

Treatments: Control = 0 mg Fe/kg BW; ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate

[FeSO4]; ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

SEM = standard error of the mean.

76

1

2

Figure 2.8. Treatment by hour (P = 0.93) and pooled hour (SEM = 12.94; P = 0.001)
responses of serum unsaturated Fe-binding capacity in Experiment 2
310
290

Serum UIBC. µg/dL

270
250
230
210
190
Control

Ferrous sulfate

Ferric sulfate

170

Hour effect pooled across treatments

150
1
1

2

3

4

5

6

Hour

7

8

9

11

12

Treatments: Control = 0 mg Fe/kg BW; ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate

[FeSO4]; ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

10

SEM = standard error of the mean.

77

1

2

Figure 2.9. Treatment by hour (P = 0.82) and pooled hour (SEM = 2.78; P < 0.0001)
3

responses of serum Fe saturation in Experiment 2
45

Serum Fe saturation, %

40

35

30

25

Control

Ferrous sulfate

Ferric sulfate

Hour effect pooled across treatments

20
1

2

3

4

5

6

7

8

9

10

11

12

Hour
1

Treatments: Control = 0 mg Fe/kg BW; ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate

[FeSO4]; ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2
3

SEM = standard error of the mean.
Fe Saturation = Fe/total Fe binding capacity.

78

1

2

Figure 2.10. Treatment by hour (P = 0.87) and pooled hour (SEM = 0.05; P < 0.0001)
responses of serum Zn concentration in Experiment 2
1
0.9

Serum Zn, µg/mL

0.8
0.7
0.6
0.5
0.4
0.3
0.2
Control

Ferrous sulfate

Ferric sulfate

0.1

Hour effect pooled across treatments

0
1

2

3

4

5

6

7

8

9

10

11

12

Hour
1

Treatments: Control = 0 mg Fe/kg BW; ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate

[FeSO4]; ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

SEM = standard error of the mean.

79

1

2

Figure 2.11. Treatment by hour (P = 0.19) and pooled hour (SEM = 0.33; P = 0.001)
responses of serum α-Tocopherol concentration in Experiment 2
7

Serum α-Tocopherol, µg/mL

6
5
4
3
2
Control

Ferrous sulfate

Ferric sulfate

1

Hour effect pooled across treatments

0
1

2

3

4

5

6

7

8

9

10

11

12

Hour
1

Treatments: Control = 0 mg Fe/kg BW; ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate

[FeSO4]; ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

SEM = standard error of the mean.

80

1

2

Figure 2.12. Treatment by hour (P = 0.33) and pooled hour (SEM = 0.07; P = 0.08) responses
of whole blood glutathione peroxidase activity in Experiment 2

Whole blood GSH-Px, IU/g Hb

2.4
2.35
2.3
2.25
2.2
2.15
2.1
2.05
2

Control
Ferric sulfate

Ferrous sulfate
Hour effect pooled across treatments

0

6

12

Hour
1

Treatments: Control = 0 mg Fe/kg BW; ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate

[FeSO4]; ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

SEM = standard error of the mean.

81

REFERENCES

82

REFERENCES

Akbulut, H., K. G. Akbulut, F. Icli, and A. Büyükcelik. 2003. Daily variations of plasma
malondialdehyde levels in patients with early breast cancer. Cancer Detect. Prev. 27:122126.
Ammerman, C. B., J. M. Wing, B. G. Dunavant, W. K. Robertson, J. P. Feaster, and L. R.
Arrington. 1967. Utilization of inorganic iron by ruminants as influenced by form of iron
and iron status of the animal. J. Anim. Sci. 26:404-410.
Beard, J. L., and H. D. Dawson. 1997. Iron. Pages 275-334 in Handbook of Nutritionally
Essential Mineral Elements. B. L. O’Dell and R. A. Sunde, eds. Marcel Dekker, New
York, NY.
Bradford, B. J., and M. S. Allen. 2005. Phlorizin administration increases hepatic gluconeogenic
enzyme mRNA abundance but not feed intake in late-lactation dairy cows. J. Nutr.
135:2206-2211.
Braughler, J. M., L. A. Duncan, and R. L. Chase. 1986. Involvement of iron in lipid
peroxidation: importance of ferric to ferrous ratio in initiation. J. Biol. Chem. 261:1028210289.
Campbell, A. G., M. R. Coup, W. H. Bishop, and D. E. Wright. 1974. Effect of elevated iron
intake on the copper status of grazing cattle. N. Z. J. Agric. Res. 17:393-399.
Chow, C. K., K. Reddy, and A. L. Tappel. 1973. Effects of dietary vitamin E on the activities of
the glutathione peroxidase system in rat tissues. J. Nutr. 103:618-624.
Colter, A., and R.L. Mahler. 2006. Iron in Drinking Water. PNW 589. Pacific Northwest
Extension Publication.
Coup, M. R., and A. G. Campbell. 1964. The effect of excessive iron intake upon the health and
production of dairy cows. N. Z. J. Agric. Res. 7: 624-638
Dove, C. R., and R. C. Ewan. 1990. Effect of excess dietary copper, iron or zinc on the
tocopherol and selenium status of growing pigs. J. Anim. Sci. 68:2406-2413.
Dutra de Oliveira, J. E., J. B. Ferreira, V. P. Vasaconcellos, and J. S. Marchini. 1994. Drinking
water as an iron carrier to control anemia in preschool children in a day-care center. J.
Am. Coll. Nutr. 13:198-202.
EPA (U.S. Environmental Protection Agency). 1994. Method 200.7. Determination of metals
and trace elements in water and wastes by inductively coupled plasma-atomic emission

83

spectrometry. Edition 4.4. Available at http://www.epa.gov/sam/pdfs/EPA-200.7.pdf.
Accessed April 20, 2011.
Ferreira, J. F., R. A. Aranda, M. P. Bianchi, I. D. Desai, and K. E. Dutra de Oliveira. 1991.
Utilization of drinking water as a vehicle for nutrients: experimental studies with iron.
Arch. Latinoam. Nutr. 41:400-408.
Furugouri, K. 1971. Normal values and physiological variations of plasma iron and total ironbinding capacity in pigs. J. Anim. Sci. 32:667-672
Gressley, T. F., S. M. Reynal, J. J. Olmos Colmenero, G. A. Broderick, and L. E. Armentano.
2006. Technical Note: Development of a tool to insert abomasal infusion lines into dairy
cows. J. Dairy Sci. 89:3965-3967.
Hansen, S. L., M. S. Ashwell, A. J. Moeser, R. S. Fry, M. D. Knutson and J. W. Spears. 2010.
High dietary iron reduces transporters involved in iron and manganese metabolism and
increases intestinal permeability in calves. J. Dairy Sci. 93:656-665.
Hashim, Z., L. Woodhouse, and J. C. King. 1996. Interindividual variation in circulating zinc
concentrations among healthy adult men and women. Int. J. Food Sci. Nutr. 47:383-390.
Ho, S. Y., W. J. Miller, R. P. Gentry, M. W. Neathery, and D. M. Blackmon. 1984. Effects of
high but nontoxic dietary manganese and iron on their metabolism by calves. J. Dairy
Sci. 67:1489-1495.
Jenkins, K. J., and M. Hidiroglou. 1987. Effect of excess iron in milk replacer on calf
performance. J. Dairy Sci. 70:2349-2354.
Jacobs, A. 1980. The pathology of iron overload. Pages 427-459 in Iron in Biochemistry and
Medicine. Vol. 2. A. Jacobs and M. Worwood, eds. Academic Press, London, UK.
Kim, Y., C. E. Carpenter, and A. W. Mahoney. 1993. Gastric acid production, iron status and
dietary phytate alter enhancement by meat of iron absorption in rats. J. Nutr. 123:940946.
Mansion, R., G. J. Rowlands, W. Little, and K. A. Collis. 1981. Variability of the blood
composition of dairy cows in relation to time of day. J. Agric. Sci. (Camb.). 96:593-598.
Markowitz, M. E., J. F. Rosen, and M. Mizruchi. 1985. Circadian variations in serum zinc (Zn)
concentrations: correlation with blood ionized calcium, serum total calcium and
phosphate in humans. Am. J. Clin. Nutr. 41:689-696.
McMillan, E. M., and D. J. F. Rowe. Clinical significance of diurnal variation in the estimation
of plasma zinc. Clin. Exp. De 1982.rmatol. 7:629-631.

84

Milchak, L. M., and J. D. Bricker. 2002. The effects of glutathione and vitamine E on iron
toxicity in isolated rat hepatocytes. Toxicol. Lett. 126:169-177.
Miller, W. J., R. P. Gentry, D. M. Blackmon, and H. H. Fosgate. 1991. Effects of high dietary
iron as ferrous carbonate on performance of young dairy calves. J. Dairy Sci. 74:19631967.
Morgan, B., and O. Lahav. 2007. The effect of pH on the kinetics of spontaneous Fe(II)
oxidation by O2 in aqeous solution – basic principles and a simple heuristic description.
Chemosphere. 68:2080-2084.
Nierenberg, D. W., and T. Stukel. 1987. Diurnal variation in plasma levels of retinol, tocopherol
and β-carotene. Am. J. Med. Sci. 294:187-190.
Nickander, K. K., J. D. Schmelzer, D. A. Rohwer, and P. A. Low. 1994. Effect of α-tocopherol
deficiency on indices of oxidative stress in normal and diabetic peripheral nerve. J.
Neurol. Sci. 126:6-14.
National Research Council. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad.
Sci., Washington, DC
Sharp, P. 2004. The molecular basis of copper and iron interactions. Proc. Nutr. Soc. 63:563-569.
Sinnah, R., J. R. Doggart, and D. W. Neill. 1969. Diurnal variations of the serum iron in normal
subjects and in patients with haemochromatosis. Br. J. Haematol. 17:351-358.
Solomans, N. W., and R. A. Jacob. 1981. Studies on the bioavailability of zinc in humans: effect
of heme and nonheme iron on the absorption of zinc. Am. J. Clin. Nutr. 34:475-482.
Sordillo, L. M., and S. L. Aitken. 2009. Impact of oxidative stress on the health and immune
function of dairy cattle. Vet. Immunol. Immunopathol. 128:104-109.
Standish, J. F., C. B. Ammerman, C. F. Simpson, F. C. Neal, and A. Z. Palmer. 1969. Influence
of graded levels of dietary iron as ferrous sulfate on performance and tissue mineral
composition of steers. J. Anim. Sci. 29:496-503.
Standish, J. F., and C. B. Ammerman. 1971. Effect of excess dietary iron as ferrous sulfate and
ferric citrate on tissue mineral composition of sheep. J. Anim. Sci. 33:481-484.
Stengle, J. M., and A. L. Schade. 1957. Diurnal-nocturnal variations of certain blood constituents
in normal human subjects: plasma iron, siderophilin, bilirubin, copper, total serum
protein and albumin, haemoglobin and haematocrit. Br. J. Haematol. 3:117-124.

85

Walker, S. E., T. W. Paton, D. H. Cowan, M. A. Manuel, and G. Dranitsaris. 1989. Bioavilability
of iron in oral ferrous sulfate preparations in healthy volunteers. Can. Med. Assoc. J.
141:543-547.
Wiltink, W. F., J. Kruithof, C. Mol, G. Bos, and H. G. Van Eijk. 1973. Diurnal and nocturnal
variations of the serum iron in normal subjects. Clin. Chim. Acta. 49:99-104.
Yamamoto, K., and E. Niki. 1988. Interaction of α-tocopherol with iron: antioxidant and
prooxidant effects of α-tocopherol in the oxidation of lipids in aqueous dispersions in the
presence of iron. Biochim. Biophyc. Acta. 958: 19-23.

86

CHAPTER 3
2+

3+

EFFECTS OF FERROUS (Fe ) OR FERRIC (Fe ) IRON CONTENT OF DRINKING
WATER ON PREFERENCE AND BEHAVIOR OF LACTATING DAIRY COWS
ABSTRACT
2+

Drinking water can contain high concentrations of Fe, mainly in the ferrous (Fe )
valence. Current recommended upper tolerable concentrations of Fe in drinking water for cattle
(0.3 mg/L) come from human recommendations for palatability, but cattle may be able to tolerate
higher concentrations. Our objective was to determine the effects that varying concentrations of
2+

3+

ferrous (Fe ) or ferric (Fe ) iron, and Fe-salt source might have on lactating dairy cows’
preference for drinking water offered as choices ad libitum. In four separate experiments, cows
were offered pairs of water treatments for 22-h periods and water intake and drinking behavior
were recorded. Experiment 1 treatments were: 0, 4, or 8 mg Fe/L from ferrous lactate
[Fe(C3H5O3)2] . Cows did not show a preference between water with 0 or 4 mg Fe/L, and water
intake was less with 8 compared with 0 or 4 mg Fe/L. Cows spent less time drinking water
containing 8 mg Fe/L. Total time spent drinking was positively correlated with water intake
when pooled across treatments. Experiment 2 treatments were: 0 or 8 mg Fe/L from either
ferrous (FeSO4) or ferric sulfate [Fe2(SO4)3]. Water intake was not different among treatments.
Experiment 3 treatments were: 0 or 8 mg Fe/L from either ferrous (FeCl2) or ferric chloride
(FeCl3). Again, cows did not show any preference among the treatments. Experiment 4
treatments were: 0 or 8 mg Fe/L from ferrous lactate [Fe(C3H5O3)2] , sulfate (FeSO4), or
chloride (FeCl2). Cows preferred water without added Fe, but did not exhibit any preference for

87

water containing any particular source of Fe with different anionic moieties. Cows spent less
time drinking and drank less frequently when offered water containing 8 mg Fe/L from ferrous
chloride compared with ferrous lactate or sulfate. Water intake was positively correlated with
both drinking duration and frequency when pooled across treatments in Experiment 4. This
research suggests that lactating dairy cows tolerate concentrations of Fe up to 4 mg/L without
reducing water intake. Additionally, preference is not dependent on Fe valence or Fe source.

88

INTRODUCTION
Preference is the degree to which one alternative is preferred to another. A dairy cow’s
preference for drinking water is an indicator of palatability that can be measured by differences
in water intake. Anecdotal information suggests that drinking water with Fe concentrations of 2
mg/L or greater may negatively impact milk production and cow health (Beede, Michigan State
University, East Lansing, MI, personal communication). These might be effects mediated one of
two ways: through changes in Fe and antioxidant statuses and oxidative stress, or through
drinking water preference. In this research we focused on drinking water preference.
The upper tolerable limit for Fe concentration in water recommended for dairy cows [0.3
mg/L (direct water analysis)] is based on human palatability. Only one study investigated the
effects of drinking water contaminated with Fe on ruminants. Horvath (1985) offered sheep
simulated acid mine water and measured water intake. Acid mine drainage is water contaminated
with byproducts of mining activity that often has a high Fe concentration (approximately 25-500
mg/L). Sheep were offered the same treatments 2 d in a row for 15 min twice daily. Treatments
concentrations of 75 mg Fe/L from ferric sulfate, or 145 mg Fe from ferric chloride had no
impact on water intake. However, it was not clear if the animals were offered water for the other
23.5 h of the day. Therefore water consumption may not have been reflective of preference in an
average setting. Additionally, Fe concentrations of treatments ranged from 75 to 145 mg Fe/L,
which are much greater than would be found typically in drinking water for dairy cows.
Palatability, as reflected by reduced DMI and subsequently ADG from supplemental Fe
in rations was characterized in calves and sheep. Calves fed rations supplemented with 750 mg
Fe/kg DM from ferrous sulfate (FeSO4) had reduced DMI and ADG (Hansen et al., 2010).
Dietary Fe concentrations of 1,600 mg Fe/kg from ferrous sulfate or ferric citrate (FeC6H5O7)
89

decreased DMI of sheep when compared with non-supplemented diets, although ferrous sulfate
had a greater negative impact on DMI than ferric citrate (Standish and Ammerman, 1971).
However, dairy cow intake of pasture irrigated with water containing a high concentration of
ferric hydroxide [Fe(OH)3] (17 mg Fe/L) was not different than that of cows on non-irrigated
pasture (Coup and Campbell, 1964).
The World Health Organization states that Fe concentrations greater than 0.3 mg/L will
affect taste preference in humans (WHO, 2006), but concentrations of 1 to 3 mg Fe/L from
anaerobic well water can be acceptable (WHO, 2003). Anaerobic water will contain mostly
2+

2+

ferrous iron (Fe ) (Colter and Mahler, 2006), and this suggests that Fe

may be more palatable

3+

than ferric iron (Fe ).
The objective of our research was to determine if water intake responses of lactating
2+

dairy cows differ when offered water with ascending concentrations of Fe , different Fe
2+

valences (Fe

3+

or Fe ) and from different Fe sources (salts). We hypothesize that as the
2+

concentration of Fe

2+

drink water with Fe

in water increases, water intake will decrease; lactating cows will prefer to
when presented with both high Fe

2+

3+

and high Fe

drinking water, but

will prefer water with no supplemental Fe (low basal Fe concentration in control water) to any
other source.

90

MATERIALS AND METHODS
The Institutional Animal Care and Use Committee at Michigan State University approved
all experimental procedures (AUF # 03/10-024-00).

Animals, Experimental Design, and Treatments
Experiment 1. Six mid-lactation Holstein cows (181 ± 44 DIM) situated in tie-stalls were
assigned randomly in a replicated 3 x 3 Latin Square involving 3 periods, 3 treatments, in 3
possible treatment pair combinations. Experimental treatments were: 0 (0Fe), 4 (4Fe), or 8mg
Fe/L (8Fe) from ferrous lactate [Fe(C3H5O3)2] added to on-site tap water (Table 3.2).
Treatments were formulated to contain 0, 4 or 8 mg Fe/L of total recoverable Fe (prepared by
acid digestion) or 0, 1 and 2 mg/L of Fe by direct metals analysis (raw water), both analyzed by
EPA method 200.7 (1994). Treatments were offered as pairs. Each cow was offered one of these
three possible pairs (0Fe and 4Fe; 0Fe and 8Fe; 4Fe and 8Fe) in each of three experimental
periods; pairs were balanced for left-right effect.
Experiment 2. Six mid-lactation Holstein cows (187 ± 44 DIM) situated in tie-stalls were
assigned randomly in a replicated 3 x 3 Latin Square design involving 3 experimental periods, 3
experimental treatments, in 3 possible treatment pair combinations. Experimental treatments
were: 0 (Control), and 8 mg Fe/L from either ferrous (ferrous sulfate; FeSO4) or ferric sulfate
(ferric sulfate; [Fe2(SO4)3]) added to on-site tap water (Table 3.3). Treatments were formulated
to contain 0 and 8 mg Fe/L of total recoverable Fe (prepared by acid digestion) or 0 and 2 mg/L
of Fe by direct metals analysis (raw water), both analyzed by EPA method 200.7 (1994).
Treatments were offered as pairs. Each cow was offered one of these three possible pairs (control

91

and ferrous sulfate; control and ferric sulfate; ferrous sulfate and ferric sulfate) in each of three
experimental periods; pairs were balanced for left-right effect
Experiment 3. Six mid-lactation Holstein cows (180 ± 16 DIM) situated in tie-stalls were
assigned randomly in a replicated 3 x 3 Latin Square design involving 3 experimental periods, 3
experimental treatments, in 3 possible treatment pair combinations. Experimental treatments
were: 0 (Control), and 8 mg Fe/L from either ferrous (ferrous chloride; FeCl2) or ferric chloride
(ferric chloride; FeCl3) added to on-site tap water (Table 3.4). Treatments were formulated to
contain 0 and 8 mg Fe/L of total recoverable Fe (prepared by acid digestion) or 0 and 2 mg/L of
Fe by direct metals analysis (raw water), both analyzed by EPA method 200.7 (1994). However,
a mistake was made in treatment formulation, and the ferrous chloride treatment actually
contained approximately 12.5 mg/L (25% more Fe). Treatments were offered as pairs. Each cow
was offered one of these three possible pairs (control and ferrous chloride; control and ferric
chloride; ferrous chloride and ferric chloride) in each of three experimental periods; pairs were
balanced for left-right effect.
Experiment 4. Twelve mid-lactation Holstein cows (135 ± 18 DIM) situated in tie-stalls
were assigned randomly in a 6 x 6 Latin Square involving 6 experimental periods, 4
experimental treatments, in 6 possible treatment pair combinations. Experimental treatments
were: 0 (control), and 8 mg Fe/L from either ferrous lactate (ferrous lactate), sulfate (ferrous
sulfate) or chloride (ferrous chloride) added to on-site tap water (Table 3.5). Treatments were
formulated to contain 0 and 8 mg Fe/L of total recoverable Fe (prepared by acid digestion) or 0
and 2 mg/L of Fe by direct metals analysis (raw water), both analyzed by EPA method 200.7
(1994). However, a mistake was made in treatment formulation, and the ferrous chloride
treatment actually contained approximately 12.5 mg/L (25% more Fe). Each cow was offered
92

one of these six possible pairs (control and ferrous lactate; control and ferrous sulfate; control
and ferrous chloride; ferrous lactate and ferrous sulfate; ferrous lactate and ferrous chloride;
ferrous sulfate and ferrous chloride) in each of six experimental periods; pairs were balanced for
left-right effect.

Water Intake Measurement and Behavior Assessment
The day that treatment water intake measurement began, water containers were removed
approximately 2 h before the experimental period commenced. Containers were scrubbed with
detergent, rinsed, and filled with treatments according to treatment assignments. Water
containers were then placed on the left and right edges of the bunks. Cows were removed from
the tie-stall and taken to the milking parlor approximately 1 h after water containers were
removed. Therefore, cows were without drinking water for 2 h before water treatments were
offered. During their absence, the remaining 50% of the ration, withheld during morning feeding,
was added to the remaining feed in the bunks. Cows returned to their tie-stalls simultaneously,
immediately after milking, to individual access to feed and two water containers. At this time
(1600 h) the period of water intake measurement and manual recording of cow drinking behavior
began.
Every 2 h containers were removed from all feed bunks at the same time, and behavior
recording was paused. Each container was skimmed with a pool skimmer to remove as much
feed as possible that was dropped into the containers by cows during the experimental period.
Containers were weighed and the weight was recorded. If the weight was less than
approximately 45 kg, the container was refilled in increments of 5 or 10 L, until the weight
reached greater than that number. Containers were then replaced in each feed bunk at the same

93

time and the behavior recording recommenced. Behavior recording ended 12 h after the
experimental period began. Containers were weighed, refilled and replaced in the bunks.
Initially, some cows pushed containers from the bunk in order to access feed. If
containers were pushed more than 20 cm from the bunk, they were pushed back into the bunk
within cow’s reach. If the cow routinely pushed the containers away, the containers were banded
to the bunks with rubber straps. Straps did not interfere with drinking. If the containers
contained less than approximately 10 L of water before the scheduled refilling time, containers
were removed from all bunks, and the containers were refilled so that containers none were ever
empty and cows were always able to drink from either container.
Containers were left in the bunks for cows to access for 22 h. They were checked at
0900 h, weighed and refilled. Containers were removed and weighed at the end of the
experimental period (1400 h).
Behavior Assessment. Drinking behaviors were recorded for the initial 12 h of the
experimental period and categorized for analyses. Categories were lapping, drinking, and ‘other’.
Drinking was defined as a cow fully or partially submerging her muzzle in water to consume
water. Lapping was defined as a cow consuming water without submerging her muzzle.

The

‘other’ category included all behaviors associated with containers when cows were not actively
consuming water. Non-consumption behavior interaction examples included splashing, sniffing
and pushing containers away. Frequency of interaction with water containers also was evaluated.
Behavior measurements were recorded to the nearest second. The beginning of one behavior
signified the end of the previous behavior, so only beginning time of behavior was recorded. If a
behavior was associated with a water container, the designated treatment identifier was recorded.

94

Other common behaviors recorded included eating, ruminating, urinating, defecating, standing
up, lying down, resting, and standing.
Preparation of Treatments
Treatments were prepared in concentrated form (100-times final treatment concentration)
in 1 L polypropylene bottles so 1 L of concentrated solution was diluted to 100 L of final
treatment solution. During preparation of final water treatments, concentrated solutions were
diluted with on-site tap water (water composition can be found in Table 3.7) until the proper
concentration, and a volume (50 L) was reached. Additional treatment information is presented
in Tables 3.2, through and 3.5.

Water containers
Water was offered in two, 75 L capacity white Rubbermaid Brute™ (Rubbermaid
Commercial Products, Winchester, VA) storage tote containers (W: 38.4 cm, x H: 44.1 cm x L:
70.8 cm) each placed in the left or right side of individual cow’s feed bunks with feed situated
between the containers. There was a vacant tie-stall between cows to avoid cow social
interactions that potentially might affect water intake.

Training
Cows were blocked from access to water in the milking parlor lane, so the only source of
water was the source provided in the tie-stall throughout the pre-experimental and experimental
periods. Nine d prior to the beginning of each experiment, cows were offered water from the onsite tap water source in the experimental water containers for 48 h to ensure that all cows became
accustomed to the containers. Containers were removed from the bunks, weighed and refilled

95

every 6 h during training. Every 24 h containers were removed, emptied, scrubbed with
detergent, rinsed well and refilled. There were no issues with cows refusing to drink from
containers. Following the initial training period cows were placed in tie-stalls with individual
automatic water cups, as in normal settings, and water intake was measured with in-line water
flow meters for 5 d to establish baseline water intake. Forty-eight h immediately prior to actual
data collection cows were moved to stalls without automatic water cups, and water was
accessible only in the experimental water containers to re-accustom the cows to the containers
and establish baseline intake from water containers. All cows were exposed to water containers
for the same length of time during pre-experimental acclimation periods and always located with
an unoccupied stall on each side.

Common Feeding, Water, and Milk Data Collection and Analyses
A basal diet formulated to meet or exceed NRC (2001) recommendations for all nutrients
(including Fe) was fed ad libitum throughout each experiment (Table 3.6). Cows were fed once
daily in their individual tie-stalls, and milked twice daily. Milk yield and water and feed intake
were recorded daily for the entire experimental period, including adjustment periods. Preexperimental feed intake, water intake, and milk yield and composition are presented in Table
3.1 for characterization of experimental animals.
Feed intakes were measured daily throughout the experiment. Individual feed
ingredients were sampled the week before each collection period began, and again between
experimental periods. Each individual feed sample was weighed and dried in a forced-air oven at
55°C for 72 h until completely dry to determine DM percentage. Dried samples were ground
through a 2 mm screen in a Wiley mill (Arthur H. Thomas, Philadelphia, PA). Similar feed

96

ingredient samples were pooled by experiment and sent for analysis by Dairy One Laboratories
(Ithaca, NY). Milk samples were taken once per experiment from one AM and PM milking each
and sent for composition analysis at Michigan DHIA Universal Laboratory Services (Lansing,
MI). Water was sampled once per period from each treatment. Similar samples were pooled by
experiment and one sample per treatment in each experiment was analyzed by each of two
separate methods for Fe. Raw water samples were sent to Midwest Laboratories (Omaha, NE)
for direct metals water analysis (Livestock water analysis, EPA 200.7 method (1994)). A second
sample was acidified (pH < 2.0) with 1 mL nitric acid, and sent to the same laboratory for acid
digestion and metals analyses (total recoverable metals analysis using EPA 200.7 method
(1994)). Water analyses are presented in Tables 3.2 through 3.5. Internal water reference
samples containing either 0, 1, 4 and 8 mg of Fe/L from ferrous lactate were placed as unknowns
within the set of samples in triplicate for analysis. The intra-assay %CV was 4.0%.

Statistical Analyses
Data were analyzed by method of least-squares ANOVA using the Mixed Models
procedure of SAS (Version 9.2, SAS Institute Inc.). Data were analyzed as a Latin Square design
with a split plot in which treatments were sub-plots within cow, period, and treatment pair whole
plots. Differences in cumulative and non-cumulative (2 h time periods) water intakes were
compared between treatments. Data of hourly measurements were analyzed using the
REPEATED statement, and the covariance structure that resulted in the lowest Bayesian
information criterion (BIC) was used. Fixed effects in the model included treatment, hour, and
treatment by hour interaction. Random effects included cow and period. If the interaction term
was non-significant (P > 0.15), intake was analyzed separately by hour. Residual distributions
97

were tested for normality using the Shapiro-Wilk test, and outliers were identified using Cook’s
D statistic. During evaluation of the full model, including interaction terms, intake data were
transformed by taking the square root to accommodate the assumptions of normality. During
analyses of individual hour water intake responses and total behavior responses, if required, data
also were transformed to either square root or natural log if required to accommodate the
assumptions of normality. Results were back transformed for presentation. Differences between
treatments were identified using the PDIFF option with Tukey’s adjustment. Least-squares
means are presented. Significance for main effects was declared at P < 0.05, and a trend toward
significance at P < 0.10. Significance for interaction effects was declared at P < 0.10, and a
trend toward significance at P < 0.15. Pearson correlation coefficients were calculated for the
relationships between water intake and behavior responses using the Correlation procedure.

98

RESULTS AND DISCUSSION
Treatment Effects on Water Intake
Experiment 1. Our objective was to examine the effects of drinking water Fe
concentration on preference. Cows were offered pairs of treatments as 0, 4 or 8 mg Fe/L supplied
as ferrous lactate in drinking water. We were concerned initially that cows would be induced to
drink water containing additional Fe because it was novel to them, however, this was not the
case. There was no interaction in either cumulative (P = 0.96) or non-cumulative water intake (P
= 0.98) over time. There was an overall effect of treatment (P < 0.05) at the end of the total
experimental period (22 h) (Table 3.8). At the end of the experimental period (22 h) intake of
4Fe tended to be greater than intake of 8Fe (P = 0.07), but not different from 0Fe. Horvath
(1985) found that sheep offered 145 mg Fe/L from ferric sulfate Fe2(SO4)3 had reduced water
intakes, but when supplemental Fe concentration was reduced to 75 mg Fe/L from the same
source, intake was not different from Control. We expected that 0Fe would be more acceptable
to cattle than 4Fe water, but 4 mg Fe/L did not impact drinking water preference. There was no
difference in total intakes based on which pair of treatments was offered. These results suggest
that the 0.3 mg Fe/L tolerable concentration for lactating dairy cows may be too severe.
However, further evaluation is warranted to evaluate the potential contribution of valence and
source of Fe in drinking water.
Experiments 2 and 3. In Experiment 2, our objective was to determine the effect of Fe
3+

or Fe

2+

on drinking water preference. Cows were offered pairs of water treatments with 0 or 8

mg Fe/L as either ferrous or ferric sulfate. There was no interaction between treatment and time
for cumulative (P = 0.97) or non-cumulative (P = 0.69) water intake over the 22 h period.
Overall, there was no difference between total intakes for any treatments offered (P = 0.28;

99

Table 3.9). However, the total intake of cows on the Control treatment was at least 10 L greater
than other treatments. Throughout the 22 h period water intakes for Control and ferric sulfate
were numerically similar, until the final measurement. There was no interaction between
treatment and hour. There also was no difference in total water intake due to pair of treatments
offered.
In Experiment 3, we evaluated the difference between the effects of Fe

2+

3+

or Fe

on

drinking water preference using a different Fe source. Cows were offered pairs of water
treatments as 0 or 8 mg Fe/L as either ferrous or ferric chloride. However, the ferrous chloride
treatment actually contained 12.5 mg Fe/L. There was no effect of treatment on total water intake
(Table 3.10). No interaction occurred between treatment and hour for cumulative (P = 1.0) or
non-cumulative (P = 0.54) water intakes. There was no impact of pair of treatments offered on
water intake. Although not statistically different, water intake for ferrous chloride was
numerically lower than the other treatments throughout the experimental period. These results
are similar to results found in Experiment 2. In both experiments intakes of waters containing
2+

Fe

were numerically lower than other treatments offered. We suspect that the absence of

significant differences was due to insufficient statistical power (approximately 49% and 40% for
Experiments 2 and 3 respectively), as the variation among cows was greater than expected, and
earlier published data were not available in the literature with variance estimates.
During both Experiments 2 and 3, it was noted that there was a change in the appearance
2+

of Fe

3+

(ferrous sulfate and ferrous chloride) and Fe
2+

treatments. Throughout the experiments, Fe

(ferric sulfate and ferric chloride)

treatment colors increased from clear to reddish-

brown due to exposure to air, presumably oxidation. The treatment waters remained cloudy
3+

throughout the study with suspended matter, presumably containing Fe. The Fe
100

treatments also

increased in color, however, the precipitate formed by Fe appeared to sink to the bottom more
instead of staying suspended as occurred with the Fe

2+

treatments. This left red sediment at the

bottom of the water containers, but the upper portion of water still relatively clear. This leads us
3+

to suspect that cows had numerically greater intakes for Fe
evenly distributed in the Fe
2+

of Fe

2+

2+

than Fe

because Fe was more

treatments. We suggest that the presumably more even distribution

had a negative effect on drinking water palatability.

Experiment 4. Iron has a strong tendency to form organic and inorganic salt complexes
(ferrous lactate, ferrous sulfate, ferric oxide) with other elements (Hem and Cropper, 1959). It is
not known if Fe salt complexes have an effect on palatability of drinking water. In Experiment 4,
our objective was to determine if cow preference for drinking water was dependent on Fe source
supplied. Cows were offered pairs of water treatments containing 0 or 8 mg Fe/L as ferrous
lactate, ferrous sulfate, or ferrous chloride. However, the ferrous chloride treatment actually
contained 12.5 mg Fe/L. There was no interaction between treatment and time for both
cumulative (P = 0.20) and non-cumulative (P = 0.58) water intakes. There was an overall effect
of treatments on total water intake (P < 0.0001; Table 3.11). Cows preferred the control water
source to every supplemental Fe water treatment. There was no difference in intake by cows
offered different sources of Fe (P > 0.15). This result is contrary to results of Horvath (1985) in
3+

which wethers offered 145 mg Fe/L of Fe

from either ferric sulfate or ferric chloride preferred

water containing ferric chloride. Digesti and Weeth (1976) also found that water palatability,
measured as total water intake, was reduced more by sodium sulfate than an equal anionic
concentration of sodium chloride (concentrations ranged from 275 to 4,400 mg/L). In our study,
the concentrations of lactate, sulfate and chloride were relatively lower, and likely did not have
nearly as much impact on palatability as the concentrations of Fe. These results also suggest that
101

other contaminants (the anion of the salt) in water do not decrease the palatability of water with
high Fe concentration.

Water Analyses
Water samples from treatments were collected throughout the experiments. They were
prepared for analysis of Fe by each of two different methods: 1) the direct metals analysis
method of raw water (without acidification and acid digestion) known as the Livestock water
analysis; and, 2) the total recoverable metals analysis by acidification of the water and acid
digestion before analysis. Midwest Laboratories (Omaha, NE) analyzed both sample
preparations for Fe by using EPA 200.7 method (1994). The concentrations determined by the
direct analysis method did not compare to the concentrations of solutions formulated and
prepared for the experiments (Tables 3.2 through 3.5, 3.7). In general, the concentration values
by direct metals analysis of Fe were 7 - 25% of the concentrations by the total recoverable metals
analysis. The direct Fe analysis is that used in the field to evaluate water quality in dairy farms
(Beede, Michigan State University, East Lansing, MI, personal communication). In our
2+

experiments cows drank less water with Fe concentrations of 8 mg Fe /L as determined by the
total recoverable metals analysis. Therefore, these results indicate that in the field, 2 mg of Fe/L
may be closer to 8 mg Fe/L and perhaps could negatively impact water intake. In short, the
amounts of Fe that we are offering to cows through high-Fe drinking water could be near 4-fold
of what we had believed previously if one is using the Fe concentration values from the direct
metals analysis (Livestock water analysis).

102

Treatment Effects on Drinking Behavior
Drinking behavior was measured for the first 12 h of the experimental periods in each
experiment. Our objective was to determine if Fe concentration, valance and source impacted
drinking water behavior durations, total time spent interacting with each treatment, and
frequency of interactions with each treatment. The main behaviors assessed were drinking,
lapping and non-consumption interactions.
Experiment 1. We hypothesized that as concentrations of Fe increased, total duration and
frequency of drinking and lapping would decrease, and time spent interacting with the container,
but not consuming water would increase. Cows spent more time drinking 0Fe and 4Fe than the
8Fe treatment (P = 0.02; Table 3.12). Water intake also was greater for both 0Fe and 4Fe water
treatments than 8Fe. There were no other treatment effects on duration of other behaviors. There
was no difference in the frequency of occurrence of any of the drinking behaviors. Total time
spent at water containers and visit frequencies were not different due to treatment.
Experiment 2. Coinciding with our hypothesis related to water intake, we suspected that
cows would spend more time drinking from the Control water than supplemental Fe treatments,
and more time drinking water of the ferrous sulfate treatment than the ferric sulfate treatment.
We also hypothesized that time spent interacting with the containers, not consuming water,
would be greatest for ferric sulfate. There was no treatment effect on duration or frequency of
drinking behaviors (Table 3.13). There also was no effect of treatment on total time spent and
frequency of visits to water containers.
Experiment 3. Similar to our hypothesis in Experiment 2, we expected that drinking
duration and frequency would be greater for ferrous chloride than ferric chloride, and greatest for
the Control treatment water. We also expected that cows would spend more time interacting
103

with, but not consuming ferric chloride water. There was no treatment effect on duration or
frequency of drinking behaviors (Table 3.14). There also was no effect of treatment on total time
spent at each water container or frequency of visits.
Experiment 4. We expected that cows would spend more time drinking and drink more
often from Control water, and there would be no differences in behaviors among Fe sources.
There was no effect of treatment on frequency or duration of lapping or non-consumption
behaviors (Table 3.15). There was an effect of treatment on drinking duration (P = 0.005) and
frequency (P = 0.03). Cows spent more time drinking from the Control, ferrous lactate and
ferrous sulfate treatments than the ferrous chloride treatment (P = 0.02), and tended to spend
more total time drinking from the ferrous lactate treatment than the ferrous chloride treatment (P
= 0.07). However, this effect may be due to the additional iron contained in the ferrous chloride
treatment. Overall, cows drank more frequently from the Control treatment than supplemental Fe
treatments. Treatment did not impact the total time spent at each container, nor the number of
visits per container during the 12 h observation period.

Correlations between Drinking Behavior and Water Intake
The relationships between various drinking behavior durations and frequencies were
investigated for individual treatments and pooled across treatments for each experiment.
Correlations were completed for 12 h total behavior durations and corresponding 12 h total
intake.
Experiment 1. Relationships and correlation coefficients between drinking behaviors
and water intake were visually compared, and were consistent across treatments (Table 3.16).
Overall, there was no correlation between lapping duration and total water intake. However,
104

when correlations were investigated by treatment, water intake of Control treatment tended to be
negatively correlated with lapping duration (r = -0.53; P < 0.10), whereas lapping duration was
positively correlated with water intake of ferrous sulfate (r = 0.65; P < 0.05) and not related for
ferric sulfate (r = 0.28; P > 0.15). Overall, drinking duration (r = 0.62; P < 0.05) and drinking
frequency (r = 0.56; P < 0.05) were positively correlated with total 12-h water intake. No
correlations were observed for lapping or non-consumption behavior frequencies or durations.
Experiment 2. The correlation coefficients and relationships between drinking behaviors
and water intake were consistent among treatments, after visual evaluation (Table 3.17). When
behaviors and intakes were pooled across treatments, both drinking duration (r = 0.70; P <0.05)
and drinking frequency (r = 0.52; P < 0.05) were positively correlated with water intake. There
were no other relationships between drinking behaviors and water intake.
Experiment 3. There were no differences in drinking behavior durations or frequencies.
However, an interesting trend was identified between treatments. Most of the correlations
between behavior and water intake were negative for the ferrous chloride treatment (Table 3.18).
This was in contrast to Control and ferric chloride treatments, in which all relationships between
behavior and water intake were positive for these treatments. The differences could be due to the
excess Fe or the valence of the ferrous chloride treatment. Overall, there were no correlations
between water intake and behavior when responses were pooled across treatments.
Experiment 4. Correlation coefficients and relationships between individual treatment
water intakes and drinking behaviors were visually examined, and were consistent across
treatments (Table 3.19). Overall there were no relationships found. Various drinking behaviors
were not correlated with treatment water intake.

105

Overall, Fe concentration and source had the most impact on drinking duration. Research
completed by Dado and Allen (1994) indicates that lactating cows spend approximately 18.5
min/d drinking and drink 4.3 L/min. In our experiments cows spent approximately 8.8 min/12 h
drinking (17.6 min/24 h) and drank 8.3 L/min. However, our estimate does not include time
spent lapping. The differences between our research and Dado and Allen (1994) can be
attributed to differences in cow water intake (110 L/d and 77.6 L/d), milk production (45 kg/d
and 33 kg/d), and the season that the experiments took place (August and January), respectively.
In all studies across all experiments the maximum difference in total time spent drinking was
only approximately 200 s (difference between drinking duration of Control and ferrous chloride
treatments in Experiment 4). However, this could translate to a decrease of 29.3 L/d. This
decrease is illustrated in the positive correlation between time spent drinking and total water
intake among all experiments.

106

CONCLUSIONS
Our results indicate that the upper tolerable concentration of Fe in drinking water for
cattle should be re-evaluated. Total recoverable Fe concentration of 4 mg/L did not affect
negatively the drinking water preference of lactating dairy cows when compared with water
2+

containing no supplemental Fe. Water intake was less with total recoverable Fe of 8 mg Fe /L.
However, as we have shown, the total recoverable concentrations of Fe in drinking water are
about 4 times greater than conventionally believed based on the typical Livestock water analysis
generally used in the field. This implies that drinking water with 2 mg Fe/L on farm, analyzed by
a standard direct analysis method, could have the same negative effects as total recoverable Fe
concentration of 8 mg/L. Additionally, cows did not exhibit a preference between Fe

2+

3+

and Fe

in water, and did not have a preference for Fe source comparing ferrous lactate, sulfate and
chloride. Drinking duration was negatively impacted by greater Fe concentration and was
dependent on source. This research suggests that potential effects of high Fe in drinking water on
production and health could be due to reduced water intake due to reduction in palatability of
drinking water.

107

APPENDIX C
FIGURES AND TABLES REFERENCED IN CHAPTER 3

108

TABLES
Table 3.1. Pre-experiment dry matter intake, milk yield and composition, and water intake of
cows in Experiment 1 through 4
Experiment
1

1-3
DMI, kg DM

±SD

2

4

±SD

23

3.2

23

3.5

Water intake, L

109.6

16.40

108.3

18.75

Milk yield, kg

47.0

5.81

45.3

7.03

Fat

3.02

0.13

3.26

0.71

True protein

2.74

0.15

2.81

0.46

Lactose

4.64

0.08

4.57

0.73

5.53

0.09

5.46

0.87

20

34.52

25.2

4.37

Milk Composition, %

SNF

3

4

SCC , 1000 ml
5

MUN , mg/dl
1
2
3
4
5

150

187.5

20.2

2.77

Milk composition results were pooled for Experiments 1, 2, and 3.
SD = standard deviation.
SNF = solids not fat.
SCC = somatic cell count.
MUN = milk urea nitrogen.

109

Table 3.2. Characteristics of water treatments in Experiment 1
Treatments

1

0Fe

4Fe

8Fe

0.0

4.0

8.0

0.01

0.83

2.04

0.40

4.04

8.47

0.0

13.6

27.23

Concentrations of constituents (mg/L)
Fe, formulated
Fe, direct analysis

2

Fe, total recoverable

3

Lactate, formulated
1

Treaments: 0Fe = 0 mg Fe/L; 4Fe = 4 mg Fe/L from ferrous lactate [Fe(C3H5O3)2]; 8Fe = 8 mg

Fe/L from ferrous lactate [Fe(C3H5O3)2].
2

Direct metals analysis of raw (without acid digestion) water by Livestock water analysis using

EPA 200.7 method (1994) by Midwest Laboratories (Omaha, NE).
3

Total recoverable metals analysis of acidified water after acid digestion using EPA 200.7

method (1994) by Midwest Laboratories (Omaha, NE).

Table 3.3. Characteristics of water treatments in Experiment 2
Treatments

1

Control

Ferrous sulfate

Ferric sulfate

0.0

8.0

8.0

0.26

1.25

0.7

1.56

8.73

7.75

0.0

13.77

20.65

Concentrations of constituents (mg/L)
Fe, formulated
Fe, direct analysis

2

Fe, total recoverable

3

Sulfate, formulated
1

Treatments: Control = 0 mg Fe/L; ferrous sulfate = 8 mg Fe/L from FeSO4; ferric sulfate = 8

mg Fe/L from Fe2(SO4)3.
2

Direct metals analysis of raw (without acid digestion) water by Livestock water analysis using

EPA 200.7 method (1994) by Midwest Laboratories (Omaha, NE).
3

Total recoverable metals analysis of acidified water after acid digestion using EPA 200.7

method (1994) by Midwest Laboratories (Omaha, NE).

110

Table 3.4. Characteristics of water treatments in Experiment 3
Treatments

1

Control

Ferrous chloride

Ferric chloride

0.0

12.5

8.0

0.06

2.04

0.3

1.19

12.1

7.02

0.0

18.79

32.12

Concentrations of constituents (mg/L)
Fe, formulated
Fe, direct analysis

2

Fe, total recoverable

3

Chloride, formulated
1

Treatments: Control = 0 mg Fe/L; ferrous chloride = 8 mg Fe/L from FeCl2; ferric chloride = 8

mg Fe/L from FeCl3.
2

Direct metals analysis of raw (without acid digestion) water by Livestock water analysis using

EPA 200.7 method (1994) by Midwest Laboratories (Omaha, NE).
3

Total recoverable metals analysis of acidified water after acid digestion using EPA 200.7

method (1994) by Midwest Laboratories (Omaha, NE).
Table 3.5. Characteristics of water treatments in Experiment 4
Treatments
Control

1

Ferrous

Ferrous

Ferrous

lactate

sulfate

chloride

0.0

8.0

8.0

12.5

0.04

1.6

1.88

1.82

0.31

7.58

7.27

11.2

Lactate, formulated

0.0

27.22

0.0

0.0

Sulfate, formulated

0.0

0.0

13.77

0.0

Chloride, formulated

0.0

0.0

0.0

18.14

Concentrations of constituents (mg/L)
Fe, formulated
Fe, direct analysis

2

Fe, total recoverable

1

3

2+

Treatments: Control = 0 mg Fe/L; ferrous lactate = 8 mg Fe /L from Fe(C3H5O3)2; ferrous
2+

2+

sulfate = 8 mg Fe /L from FeSO4; ferrous chloride = 8 mg Fe /L from FeCl2.
2

Direct metals analysis of raw (without acid digestion) water by Livestock water analysis using

EPA 200.7 method (1994) by Midwest Laboratories (Omaha, NE).
3

Total recoverable metals analysis of acidified water after acid digestion using EPA 200.7

method (1994) by Midwest Laboratories (Omaha, NE).
111

Table 3.6. Ingredient and nutrient composition of diets fed in Experiments 1 through 4
Experiment
1

1-3
Ingredient

4

------------------Percent of dietary DM----------------

Chopped alfalfa hay

3.74

3.93

35.63

33.61

Whole cottonseed

5.70

6.20

Energy booster

0.40

0.42

Grass hay

4.83

5.17

Ground corn

9.52

10.09

Alfalfa haylage

9.86

8.89

High moisture corn

6.98

7.41

14.64

15.19

5.25

5.55

3.27

3.45

49.40

46.60

Corn silage

Soybean meal, 48% CP
Soyhulls
Mineral-vitamin mix

2

Nutrient Composition
DM, %

-----------------Percent of dietary DM--------------NDF

32.90

34.50

ADF

22.20

22.60

CP

17.84

17.95

Ca

1.11

1.06

P

0.35

0.37

K

1.26

1.34

Na

0.31

0.35

S

0.22

0.37

Mg

0.24

0.25

-------------------------mg/kg-------------------------Co

0.11

0.12

Cu

10.1

10.9

Fe

257

179

Mo

1.8

112

2.2

Table 3.6. (cont’d)
Experiment
1

1-3
Nutrient Composition

4

-------------------------mg/kg--------------------------

Mn

2

46

Zn
1

42
41

42

Feed ingredients were pooled for experiments 1-3 for the values reported in this table.
Mineral-vitamin mix contained 47.5% limestone, 22.5% sodium bicarbonate, 10.1% urea

(45%N), 8.2% magnesium sulfate, 7.5% sodium chloride, 1.6% biotin premix (1.4 g/kg), 1.3%
trace mineral premix, 30 KIU/kg vitamin A, 8 KIU/kg vitamin D, 56 KIU/kg vitamin E and 0.3%
selenium yeast.

113

Table 3.7. Water quality constituent analysis of Control drinking water in Experiments 1 though
4
Experiment
1

2

3

4

Caution level

377

462

411

387

1,000

Ca

86

114

99

92

150

Cl

21

15

25

16

500

ND

ND

ND

0.3

0.07

0.26

0.06

0.04

0.3

0.4

1.56

1.19

0.31

-

Mg

30

38

31

31

Mn

0.06

0.75

0.06

0.03

NO3-N

ND

ND

ND

ND

25

Na

9

7

11

7

150

SO4

53

122

59

61

300

Zn

0.17

0.14

0.06

0.14

-

Conductivity, mmhos/cm

0.58

0.71

0.63

.59

1.5

pH

7.8

7.7

7.7

7.6

1

6.5-9.0

Quality Constituent
TDS

2

Cu

ND

Fe (direct Fe analysis)
Fe (total recoverable)

1
2
3
4

4

5

3

80
-

Caution levels from Midwest Laboratories (Omaha, NE)
TDS = total dissolved solids
ND = not detected
Direct metals analysis of raw (without acid digestion) water by Livestock water analysis using

EPA 200.7 method (1994) by Midwest Laboratories (Omaha, NE).
5

Total recoverable metals analysis of acidified water after acid digestion using EPA 200.7

method (1994) by Midwest Laboratories (Omaha, NE).

114

Table 3.8. Cumulative water intake (L) by hour of cows in Experiment 1
Treatments

1
2

0Fe

4Fe

8Fe

SEM

2

10.9

10.5

8.3

1.61

6

24.5

21.8

17.9

2.31

12

35.7

35.6

26.9

3.22

Hour

17
22
a-b
1

a

47.3

ab

60.4

ab

b

45.5

33.7

a

b

61.9

46.1

4.33
4.80

Means within a row with different superscripts tended to differ (P < 0.10).

Treaments: 0Fe = 0mg Fe; 4Fe= 4 mg Fe/L from ferrous lactate [Fe(C3H5O3)2]; 8Fe = 8 mg

Fe/L from Fe(C3H5O3)2.
2

SEM = standard error of the mean.

Table 3.9. Cumulative water intake (L) by hour of cows in Experiment 2
Treatments

1
2

Control

Ferrous sulfate

Ferric sulfate

SEM

2

10.5

11.2

13.2

2.15

6

23.9

19.2

26.1

3.89

12

38.3

29.2

36.3

6.00

17

47.1

34.0

41.6

6.91

22

68.1

50.6

56.1

7.78

Hour

1

Treatments: Control = 0 mg Fe/L; ferrous sulfate = 8 mg Fe/L from FeSO4; ferric sulfate = 8

mg Fe/L from Fe2(SO4)3.
2

SEM = standard error of the mean.

115

Table 3.10. Cumulative water intake (L) by hour of cows in Experiment 3
Treatments

1
2

Control

Ferrous chloride

Ferric chloride

SEM

2

10.8

7.8

13.6

2.03

6

24.8

16.2

26.9

4.15

12

36.8

28.5

38.0

5.80

17

43.7

34.2

42.5

6.43

22

63.8

47.4

67.3

9.35

Hour

1

Treatments: Control = 0 mg Fe/L; ferrous chloride = 8 mg Fe/L from FeCl2; ferric chloride = 8

mg Fe/L from FeCl3.
2

SEM = standard error of the mean.

Table 3.11. Cumulative water intake (L) by hour of cows in Experiment 4
Treatments
Control

Ferrous lactate

1

Ferrous sulfate

Ferrous chloride

2

SEM

Hour
3

a

2

11.6

6

32.3

12

49.8

17

55.1

22

71.9

a-b
1

a
a
a
a

ab

ab

7.3

b

7.3

b

6.8

b

22.5

b

19.4

b

17.3

b

34.5

b

31.3

b

27.3

b

39.7

b

35.5

b

32.7

b

53.3

b

45.8

43.0

1.36
2.69
3.66
4.16
5.11

Means within a row with different superscripts differ (P < 0.05).
2+

Treatments: Control = 0 mg Fe/L; ferrous lactate= 8 mg Fe /L from Fe(C3H5O3)2; ferrous
2+

2+

sulfate = 8 mg Fe /L from FeSO4; ferrous chloride = 8 mg Fe /L from FeCl2.
2
3

SEM = standard error of the mean.
Hour 2 data were square root transformed for analysis, and LS means and SEM were back-

transformed for presentation in this table.

116

Table 3.12. Cumulative drinking behavior durations and frequencies of cows in Experiment 1
Treatments
0Fe

4Fe

1

8Fe

2

SEM

P-value

Behavior Duration (s)
a

a

b

Drinking

210

214

124

28.7

0.01

Lapping

196

226

234

84.8

0.65

30

28

39

18.8

0.76

Drinking

14.3

13.8

11.7

1.49

0.17

Lapping

14.7

15.6

16.3

2.63

0.57

8.8

8.1

8.5

.96

0.56

Other

3, 4

Behavior Frequency

Other
a-b
1

3

Means within a row with different superscripts differ (P < 0.05).

Treaments: 0Fe = 0 mg Fe; 4Fe= 4 mg Fe/L from ferrous lactate [Fe(C3H5O3)2]; 8Fe = 8 mg

Fe/L from Fe(C3H5O3)2.
2
3
4

SEM = standard error of the mean.
Other = interacting with water container but not consuming water.
Data were square root transformed for analysis and LS means and SEM were back-transformed

for presentation in this table.

117

Table 3.13. Cumulative drinking behavior durations and frequencies of cows in Experiment 2
Treatments

1
2

Control

Ferrous sulfate

Ferric sulfate

SEM

P-value

182

119

156

50.5

0.558

210

187

182

71.7

0.82

14

15

15

8.0

0.99

Drinking

15.4

13.1

14.6

2.17

0.58

Lapping

13.8

14.2

14.1

2.41

0.95

7.8

7.9

7.6

0.68

0.91

Behavior Duration (s)
Drinking

3

Lapping
Other

4, 5

Behavior Frequency

Other
1

4, 5

Treatments: Control = 0 mg Fe/L; ferrous sulfate = 8 mg Fe/L from FeSO4; ferric sulfate = 8

mg Fe/L from Fe2(SO4)3.
2
3

SEM = standard error of the mean.
Data were natural log transformed for analyses, and means back-transformed for presentation in

this table.
4
5

Other = interacting with water container but not consuming water.
Data were square root transformed for analysis and LS means and SEM were back-transformed

for presentation in this table.

118

Table 3.14. Cumulative drinking behavior durations and frequencies of cows in Experiment 3
Treatments

1
2

Control

Ferrous chloride

Ferric chloride

SEM

P-value

221

96

199

69.6

0.22

131

127

194

66.5

0.67

Behavior Duration (s)
Drinking
Lapping
Other

3

3

3, 4

†

46

19

56

30.9

0.09

Drinking

23.5

19.9

31.3

5.81

0.21

Lapping

22.6

22.8

27.6

3.61

0.40

9.0

9.1

13.7

2.8

0.16

Behavior Frequency

Other
1

4

Treatments: Control = 0 mg Fe/L; Ferrous chloride = 8 mg Fe/L from FeCl2; ferric Chloride = 8

mg Fe/L from FeCl3.
2
3

SEM = standard error of the mean.
Data were natural log transformed for analyses, and means back-transformed for presentation in

this table.
4
†

Other = interacting with water container but not consuming water.
Tended to be lower than ferric chloride (P < 0.09).

119

Table 3.15. Cumulative drinking behavior durations and frequencies of cows in Experiment 4
Treatments
Control

1

Ferrous

Ferrous

Ferrous

lactate

sulfate

2

SEM

P-value

chloride
49.5

.004

Behavior Duration (s)
Drinking
Lapping
Other

3

a

316

4

ab

ab

222

b†

200

106

213

205

291

257

67.1

.37

143

3, 5

123.6

113

124

35.1

0.93

6.7

2.15

0.03

Behavior Frequency
a

Drinking
Lapping
Other
a-b
1

11.2

4

ab

ab

8.7

b

8.6

8.4

8.8

9.8

9.7

2.11

.80

4.1

4, 5

3.9

3.9

3.9

.91

.99

Means within a row with different superscripts differ (P < 0.05).
2+

Treatments: Control = 0 mg Fe/L; ferrous lactate= 8 mg Fe /L from Fe(C3H5O3)2; ferrous
2+

2+

sulfate = 8 mg Fe /L from FeSO4; ferrous chloride = 8 mg Fe /L from FeCl2.
2
3

SEM = standard error of the mean.
Data were natural log transformed for analyses, and means back-transformed for presentation in

this table.
4

Data were square root transformed for analysis and LS means and SEM were back-transformed

for presentation in this table.
5
†

Other = interacting with water container but not consuming water.
Tended to differ (P < 0.10) from ferrous lactate.

120

Table 3.16. Pearson correlation coefficients for the relationships between drinking behavior and
12 h total water intake in Experiment 1
Experiment
Behavior * Water Intake

0Fe

4Fe

8Fe

Pooled across
treatments

Drinking
†

Duration

0.65371*

0.51433

0.47432

0.62254*

Frequency

0.74011*

0.34511

0.31158

0.55553*

0.64785*

0.27885

0.08379

Lapping
Duration
Frequency

-0.53377

†

†

-0.01356

0.56509

0.26319

0.22740

Duration

0.31802

-0.02181

-0.11572

0.06227

Frequency

0.49544

0.00658

-0.18433

0.14325

Other

†

P < 0.10.

*P < 0.05.

121

Table 3.17. Pearson correlation coefficients for the relationships between drinking behavior and
12 h total water intake in Experiment 2
Experiment
Behavior * Water Intake

Control

Ferrous sulfate

Ferric sulfate

Pooled across
treatments

Drinking
†

Duration

0.68347*

0.56967

0.84163*

0.70464*

Frequency

0.41922

0.66291*

0.58211*

0.52143*

Duration

0.43651

0.25246

0.31928

0.31817

Frequency

0.62491*

0.08924

0.35114

0.28454

Duration

-0.08649

-0.35700

-0.10997

-0.15631

Frequency

-0.11826

-0.11070

0.13946

-0.06315

Lapping

Other

†

P < 0.10.

*P < 0.05.

122

Table 3.18. Pearson correlation coefficients for the relationships between drinking behavior and
12 h total water intake in Experiment 3
Experiment
Behavior * Water Intake

Control

Ferrous

Ferric

Pooled across

chloride

chloride

treatments

0.38998

0.52901

Drinking
Duration
Frequency

0.49667

†
†

0.42773

0.48297

-0.07224

0.54007

0.41004

Duration

0.29902

-0.21694

0.61865*

0.35768

Frequency

0.43217

-0.46821

0.62511*

0.33905

0.76286*

-0.36625

0.58239

Lapping

Other
Duration
Frequency
†

0.53779

†

-0.40534

P < 0.10.

*P < 0.05.

123

†

0.56597

†

0.38706
0.34541

Table 3.19. Pearson correlation coefficients for the relationships between drinking behavior and
12 h total water intake in Experiment 4
Treatment
Behavior * Water Intake

Control

Ferrous

Ferrous

Ferrous

Pooled across

lactate

sulfate

chloride

treatments

0.42433*

0.42816*

0.42142

Drinking
Duration
Frequency

0.45451*

†

0.32436

†

0.39526*

0.48484*

0.43335*

0.32048

0.41149

0.10954

0.05615

0.22615

0.38021*

0.11925

Lapping
Duration
Frequency

-0.03270

0.12386

0.08579

Duration

0.27511

0.26315

Frequency

0.22752

0.03329

†

0.30659

0.06866

0.28772

0.25306

0.28659

0.09249

0.39405*

0.18444

Other

†

P < 0.10.

*P < 0.05.

124

†

REFERENCES

125

REFERENCES

Colter, A., and R.L. Mahler. 2006. Iron in Drinking Water. PNW 589. Pacific Northwest
Extension Publication.
Coup, M. R., and A. G. Campbell. 1964. The effect of excessive iron intake upon the health and
production of dairy cows. N. Z. J. Agric. Res. 7: 624-638.
Dado, R. G., and M. S. Allen. 1994. Variation in and relationships among feeding, chewing, and
drinking variables for lactating dairy cows. J. Dairy Sci. 77:132-144.
Digesti, R. D., and H. J. Weeth. 1976. A defensible maximum for inorganic sulfate in drinking
water of cattle. J. Anim. Sci. 42:1498-1502.
EPA (U.S. Environmental Protection Agency). 1994. Method 200.7. Determination of metals
and trace elements in water and wastes by inductively coupled plasma-atomic emission
spectrometry. Edition 4.4. Available at http://www.epa.gov/sam/pdfs/EPA-200.7.pdf.
Accessed April 20, 2011.
Hansen, S. L., M. S. Ashwell, A. J. Moeser, R. S. Fry, M. D. Knutson, and J. W. Spears. 2010.
High dietary iron reduces transporters involved in iron and manganese metabolism and
increases intestinal permeability in calves. J. Dairy Sci. 93:656-665.
Hem, J. D., and W. H. Cropper. 1959. Chemistry of iron in natural water: survey of ferrous-ferric
chemical equilibria and redox potentials. U. S. Geol. Surv. Water-Supply Pap. 1459-A.
Horvath, D. J. 1985. Consumption of simulated acid mine water by sheep. J. Anim. Sci. 61:474479.
National Research Council. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad.
Sci., Washington, DC.
Standish, J. F., and C. B. Ammerman. 1971. Effect of excess dietary iron as ferrous sulfate and
ferric citrate on tissue mineral composition of sheep. J. Anim. Sci. 33:481-484.
World Health Organization. 2003. Iron in drinking-water. Background document for
development of WHO guidelines for drinking-water quality. World Health Organization,
Geneva.
World Health Organization. 2006. Guidelines for drinking water quality. Vol. 1, 3rd Ed. World
Health Organization, Geneva.

126

CHAPTER 4
CONCLUSIONS AND IMPLICATIONS
The majority of Fe naturally occurring in drinking water drawn from underground wells
2+

is in the more bioavailable ferrous (Fe ) valence. Consumption of water with high Fe
concentrations could negatively impact cow health and productivity, as anecdotal information
suggests. We hypothesized that drinking water with greater than normal Fe concentrations could
affect cows in two ways; through changes in Fe status, trace mineral interactions, and increases
in oxidative stress; and/or, through decreased water intake due to cow preference. Our objectives
were to test each of these possibilities through a series of experiments.
In chapter 2, we presented results from two different experiments designed to investigate
the short-term hourly effects (0 – 12 h) of dosing in the abomasum increasing Fe concentrations
of each valence on short-term Fe status, antioxidant status, and measurements of oxidative stress.
Overall, administration to lactating dairy cows with 0, 4, or 8 mg Fe/kg BW from ferrous lactate;
or, 0 or 8 mg Fe/kg BW from either ferrous or ferric sulfate had minimal impacts on
measurements of Fe status. In the first experiment, TIBC tended to be greater for cows given
0.75 mg Fe/kg BW, and in the second experiment, 1.5 mg Fe/kg BW from ferric sulfate tended
to reduce serum Fe concentrations. No other treatment effects on any other measurements of Fe
status were detected.
Hepatic Cu concentrations declined in response to 0.75 mg Fe/kg BW administered as
ferrous lactate. Additionally there was a decrease in serum Cu concentrations in response to 1.5
mg Fe/kg BW from ferrous sulfate. These results suggest that the amount of Fe consumed by
cows from drinking water under farm conditions could potentially cause an interaction with Cu
metabolism. Campbell et al. (1974) reported that intraruminally-dosing dairy cows with high

127

concentrations (210 mg Fe/kg BW) of Fe leads to Cu deficiency. We suggest that Cu deficiency
may still develop in dairy cows consuming drinking water contaminated by Fe over a longer
period of time.
Overall results of the first two experiments suggest that at the amounts and
concentrations tested, Fe contained in water does not impact short-term Fe status. However, we
suggest that after prolonged consumption, high Fe water may negatively impact cow health
similar to high dietary Fe. We suspect that continual exposure of dairy cows to Fe from water
could decrease DMI, and increase oxidative stress and intestinal permeability, similar to the
effects that high dietary Fe

2+

in calves (Hansen, et al., 2010). Prolonged exposure also could

lead to Cu deficiency. Further research should investigate the effects of chronic Fe consumption
from drinking water on Fe metabolism, and cow health and productivity long-term.
In Chapter 3, we presented results from four drinking water preference experiments
evaluating the effects of Fe concentration, valence and source on water intake and drinking
behavior of lactating dairy cows. When cows were offered 0, 4, or 8 mg Fe/L from ferrous
lactate, there was no difference in water intake among cows offered 0 and 4 mg/L, but water
intake of cows offered the 8 mg Fe/L treatment was lower than either 0 or 4 mg/L. Intake of
2+

water containing 8 mg Fe /L was least presumably due to reduced palatability.
In two separate experiments, 8 mg Fe/L from either a Fe

2+

3+

or Fe

source did not affect

cow preference for drinking water as measured by water intake. However, we wonder if the lack
of differences may have been due to insufficient statistical power due to larger than anticipated
variance, as there were numerical decreases in both experiments in water intake of cows offered
2+

the Fe

3+

versus the Control and Fe

treatments. During the experimental period, observers
2+

noticed that a red precipitate formed in the containers with water of both Fe
128

3+

and Fe

treatments. However, in the Fe
whereas the precipitant with Fe

2+

3+

treatments, the precipitant remained suspended in the water,
treatments sank to the bottom of the containers leaving the

upper portion of water relatively clear. We believe that the numerical differences in water intake
were due to the presumably more even distribution of Fe in the treatment water with Fe

2+

3+

compared with Fe . Overall, we believe that there is not much difference in cow preference
between Fe

2+

3+

and Fe

of equal concentrations in drinking water.
2+

We suggested that 8 mg Fe /L in drinking water has a negative effect on water intake
based on results from the initial experiment described in Chapter 3. This idea is supported by
results from the final experiment described in Chapter 3. The objective was to determine if Fe
source affected water intake. Twelve cows were offered pairs of treatments: 0, or 8 mg Fe/L as
ferrous lactate, ferrous sulfate or ferrous chloride. Eight mg Fe/L reduced water intake from all
three sources, but there were no differences among waters offered with different Fe sources.
2+

This indicates that 8 mg Fe /L has a negative impact on water intake. The long-term impact of
water with high Fe concentrations on cow health and productivity may be due to a decrease in
palatability and preference. Nonetheless, the sources of Fe (or Fe salt compound) tested did not
affect preference for water with high Fe concentrations.
The currently recommended upper tolerable concentration for Fe in drinking water for
cattle is 0.3 mg/L. However, the NRC (2005) suggests that cattle may be able to tolerate higher
concentrations. In our research, Fe concentration of 4 mg/L did not affect total water intake or
drinking behavior of lactating dairy cows. These results suggest that the current upper tolerable
concentration for cattle may be lower than practically needed, in agreement with the suggestion
of NRC (2005).

129

It is important to note that the water treatments offered in our preference experiments
2+

containing 0, 4, and 8 mg Fe /L (total recoverable Fe in which acid digestion of samples
proceeds detection), corresponded to Fe concentrations of approximately 0, 1, and 2 mg Fe/L
according to the direct metal analysis (Livestock water analysis) of Midwest Laboratories,
(Omaha, NE). We found that the amount of Fe in raw water (prepared for analysis without
acidification and acid digestion) was approximately 7 to 25% the concentration of total
recoverable Fe. Both sample types were analyzed by the EPA 200.7 detection method (1994)
after preparation. Therefore, in our research we did not test conditions similar to the field
circumstances in which Fe concentrations in water were in the 4 to 10 mg/L range by the
Livestock water analysis method (direct metal analysis), which may in fact be translate to total
recoverable Fe concentrations as great as 16 to 40 mg/L.
Overall, our research suggests that the current caution level (0.3 mg/L) for Fe in drinking
water for dairy cows may be lower than practically needed, and that cattle may be able to tolerate
greater concentrations. In our research 4 mg Fe/L (total recoverable Fe) or 1 mg Fe/L (direct
water analysis) from Fe

2+

did not impact dairy cow preference for drinking water. Further

research needs to be completed to determine if prolonged exposure and consumption of water
with 4 mg total recoverable Fe/L or greater has negative impacts on cow health.

130

REFERENCES

131

REFERENCES
Campbell, A. G., M. R. Coup, W. H. Bishop, and D. E. Wright. 1974. Effect of elevated iron
intake on the copper status of grazing cattle. N. Z. J. Agric. Res. 17:393-399.
EPA (U.S. Environmental Protection Agency). 1994. Method 200.7. Determination of metals
and trace elements in water and wastes by inductively coupled plasma-atomic emission
spectrometry. Edition 4.4. Available at http://www.epa.gov/sam/pdfs/EPA-200.7.pdf.
Accessed April 20, 2011.
Hansen, S. L., M. S. Ashwell, A. J. Moeser, R. S. Fry, M. D. Knutson and J. W. Spears. 2010.
High dietary iron reduces transporters involved in iron and manganese metabolism and
increases intestinal permeability in calves. J. Dairy Sci. 93:656-665.
National Research Council. 2005. Mineral Tolerance of Animals. 2nd rev. ed. Natl. Acad. Sci.,
Washington, DC.

132

APPENDIX D
ADDITIONAL FIGURES FROM CHAPTER 2

133

FIGURES
1

2

Figure 4.1 Treatment by hour (P = 0.34) and pooled hour (SEM = 19.61; P = 0.17) responses
3

of serum total Fe binding capacity in Experiment 1
550

Serum TIBC, µg/dL

500

450

400

350
0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

300
1
1

2

3

4

5

6

7
Hour

8

3

10

11

12

Treatments: 0Fe = 0 mg Fe; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate [Fe(C3H5O3)2];

1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

9

SEM = standard error of the mean.
Total Fe binding capacity = Fe + unsaturated Fe binding capacity.

134

1

2

Figure 4.2. Treatment by hour (P = 0.88) and pooled hour (SEM = 11.43; P = 0.21) responses
3

of serum total Fe binding capacity in Experiment 2
550

Serum TIBC. µg/dL

500

450

400

350

Control

Ferrous sulfate

Ferric sulfate

Hour effect pooled across treatments

300
1

2

3

4

5

6

7

8

9

10

11

12

Hour
1

Treatments: Control = no Fe; Ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate [FeSO4];

Ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2
3

SEM = standard error of the mean.
Total Fe binding capacity = Fe + unsaturated Fe binding capacity.

135

1

2

Figure 4.3. Treatment by hour (P = 0.63) and pooled hour (SEM = 22.37, P = 0.14) responses
of serum unsaturated Fe binding capacity in Experiment 1
350
330

Serum UIBC, µg/dL

310
290
270
250
230
210
190

0Fe

170

0.75Fe

1.5Fe

Hour effect pooled across treatments

150

1
1

2

3

4

5

6

Hour	
  

8

9

10

11

12

Treatments: 0Fe = 0 mg Fe; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate [Fe(C3H5O3)2];

1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

7

SEM = standard error of the mean.

136

1

2

Figure 4.4. Treatment by hour (P = 0.99) and pooled hour (SEM = 0.06; P = 0.37) responses
of serum Cu concentration in Experiment 1
1.2

Serum Cu, µg/mL

1
0.8
0.6
0.4
0.2

0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

0
1

2

3

4

5

6

7

8

9

10

11

12

Hour
1

Treatments: 0Fe = 0 mg Fe; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate [Fe(C3H5O3)2];

1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

SEM = standard error of the mean.

137

1

2

Figure 4.5. Treatment by hour (P = 0.61) and pooled hour (SEM = 0.06; P = 0.82) responses
of serum Cu concentration in Experiment 2
1.2

Serum Cu µg/mL

1

0.8

0.6

0.4

0.2

Control

Ferrous sulfate

Ferric sulfate

Hour effect pooled across treatments

0
1

2

3

4

5

6

7

8

9

10

11

12

Hour
1

Treatments: Control = no Fe; Ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate [FeSO4];

Ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

SEM = standard error of the mean.

138

1

2

Figure 4.6. Treatment by hour (P = 0.21) and pooled hour (SEM = 0.34; P = 0.36) responses
of serum α-Tocopherol concentration in Experiment 1
7

Serum α-Tocopherol, µg/mL

6
5
4
3
2
1

0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

0
1
1

2

3

4

5

6

Hour

8

9

10

11

12

Treatments: 0Fe = 0 mg Fe; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate [Fe(C3H5O3)2];

1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

7

SEM = standard error of the mean.

139

1

2

Figure 4.7. Treatment by hour (P = 0.97) and pooled hour (SEM = 0.09; P = 0.21) responses
of whole blood glutathione peroxidase activity in Experiment 1
2.7

GSH-Px activity, IU/g Hb

2.6
2.5
2.4
2.3
2.2
2.1

0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

2
0
1

6
Hour

Treatments: 0Fe = 0 mg Fe; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate [Fe(C3H5O3)2];

1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

12

SEM = standard error of the mean.

140

1

2

Figure 4.8. Treatment by hour (P = 0.69) and pooled hour (SEM = 21.69; P = 0.45) responses
of hepatic Fe concentration in Experiment 1
400

Hepatic Fe, µg/g dry tissue

350
300
250
200
150
0Fe

0.75Fe

1.5Fe

Hour effect pooled across treatments

100
0
1

18
Hour

Treatments: 0Fe = 0 mg Fe; 0.75Fe = 0.75 mg Fe/kg BW from ferrous lactate [Fe(C3H5O3)2];

1.5Fe = 1.5 mg Fe/kg BW from ferrous lactate.
2

36

SEM = standard error of the mean.

141

1

2

Figure 4.9. Treatment by hour (P = 0.69) and pooled hour 2 (SEM = 30.04; P = 0.57)
responses of hepatic Fe concentration in Experiment
400

Hepatic Fe, µg/g dry tissue

350

300

250

200
Control

Ferrous sulfate

Ferric sulfate

150

Hour effect pooled across treatments

100
0

18

36

Hour
1

Treatments: Control = no Fe; Ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate [FeSO4];

Ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

SEM = standard error of the mean.

142

1

2

Figure 4.10. Treatment by hour (P = 0.57) and pooled hour (SEM = 25.54; P = 0.86) responses
of hepatic Cu concentration in Experiment 2
800

Hepatic Cu, µg/g dry tissue

700
600
500
400
Control

Ferrous sulfate

Ferric sulfate

300

Hour effect pooled across treatments

200
0

18

36

Hour
1

Treatments: Control = no Fe; Ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate [FeSO4];

Ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

SEM = standard error of the mean.

143

1

2

Figure 4.11. Treatment by hour (P = 0.57) and pooled hour (SEM = 9.50; P = 0.38) responses
of hepatic Zn concentration in Experiment 2
180

Hepatic Zn, µg/g dry tissue

170
160
150
140
130
120
110
100
90

Control

Ferrous sulfate

Ferric sulfate

Hour effect pooled across treatments

80
0
1

18
Hour

Treatments: Control = no Fe; Ferrous sulfate = 1.5 mg Fe/kg BW from ferrous sulfate [FeSO4];

Ferric sulfate = 1.5 mg Fe/kg BW from ferric sulfate [Fe2(SO4)3].
2

36

SEM = standard error of the mean.

144

APPENDIX E
DEVELOPMENT OF STUDY PROTOCOL TO MEASURE DRINKING WATER
PREFERENCE AMONG LACTATING DAIRY COWS

145

DEVELOPMENT OF STUDY PROTOCOL TO MEASURE DRINKING WATER
PREFERENCE AMONG LACTATING DAIRY COWS
ABSTRACT
This work was undertaken to develop methods for subsequent drinking water preference
experiments to evaluate the preferences of dairy cows for various drinking water treatments with
different concentrations and sources of Fe. The objective was to develop a procedure and
temporal sequence as a study protocol to measure water intake and animal behavior among
lactating dairy cows given two treatment choices simultaneously. This experiment was used to
determine the optimal time to withhold drinking water before offering different water treatments
to determine preference, and the length of time to measure water intake and animal behavior
when treatments are offered. In protocol development drinking water with a high concentration
of NaCl (2%), generally disliked by cows, and Control water from the on-site tap source were
employed to create definitive preference choice. Cows were withheld from water access for 2 h
during afternoon milking and treatment preparation. They returned from the parlor to both water
treatments, and the experimental period began. Both treatments were offered to each cow each
period, for three periods. Periods were designed originally to last for 12 h but were extended to
22 h. Water intakes were measured every 2 h for 12 h and then at 17 and 22 h, and drinking
behavior was recorded for the initial 12 h of each period. Cows initially preferred NaCl water to
tap water, but approximately 8 h into the first period, the control water was preferred, and
continued to be preferred through the remainder of the experiment. Twenty-two hours was
adequate for cows to establish preference for a water treatment. Cows had adequate thirst drive
to establish preference through water intake, and cows tested both treatments. A method was
developed to observe and manually record cow drinking behavior that allowed us to successfully

146

identify differences in behavior as affected by treatment. Cows spent more time drinking from
Control water than 2% NaCl, and more time lapping from NaCl than Control.

147

INTRODUCTION
Researchers have used two-choice preference testing in experiments with cattle and other
species (Goatcher and Church, 1970a, b; Coppock et al., 1974). In general, two-choice
preference testing is used to determine differences in palatability of feed ingredients and
supplements. However, Goatcher and Church (1970a, b) completed a series of studies to
determine the taste responses of ruminants to various stimuli using water as a carrier. They
evaluated flavors including sucrose, NaCl, and acids. These experiments determined the upper
and lower discrimination thresholds, and preference and rejection thresholds for different tastes
using weaned Holstein heifer calves and other ruminants. The researchers were not concerned
about the intake response of animals to individual flavors in water, but to the flavors themselves.
They also offered tastes that are not usually found in water.
Additionally, Goatcher and Church (1970b) utilized weaned heifer calves in their
experiments that require much less water as a function of bodyweight on a daily basis compared
to high-producing lactating dairy cows. Although the researchers were successful in evaluating
taste responses in their experiments, we wanted to tailor our experiments more directly to our
objectives and experimental animal. In order to properly test our hypotheses in future
experiments, we wanted to ensure that the length of experimental periods and water withholding
period would allow us to successfully determine differences in total water intake between
treatments for lactating dairy cows.
Therefore, the aim of this work was to determine the optimal experimental period length
and water withholding period, and to develop a protocol for measuring drinking behavior and
evaluate cow preference for different water treatments.

148

MATERIALS AND METHODS
The Institutional Animal Care and Use Committee at Michigan State University approved
all experimental procedures (AUF # 03/10-024-00).

Animals, Experimental Design, and Treatments.
Four mid-lactation Holstein cows (134 ± 22 DIM) were assigned in a completely
randomized design. Experimental treatments were: drinking water from the on-site tap source
(Control; Table 5.3), or water from the same on-site source containing 2% NaCl (NaCl).
Treatments were offered as a pair. Each cow was offered both treatments in each of three
experimental periods, and pairs were balanced for left-right effect.

Water Intake Measurement and Behavior Assessment
The day that treatment water intake measurement began, water containers were removed
approximately 2 h before the experimental period commenced. Containers were scrubbed with
detergent, rinsed, and filled with treatments according to treatment assignments. Water
containers were then placed on the left and right edges of the bunks. Cows were removed from
the tie-stall and taken to the milking parlor approximately 1 h after water containers were
removed. Therefore, cows were without drinking water for 2 h before water treatments were
offered. During their absence, the remaining 50% of the ration, withheld during morning feeding,
was added to the remaining feed in the bunks. Cows returned to their tie-stalls simultaneously,
immediately after milking, to individual access to feed and two water containers. At this time
(1600 h) the period of water intake measurement and manual recording of cow drinking behavior
began.

149

Every 2 h containers were removed from all feed bunks at the same time, and behavior
recording was paused. Each container was skimmed with a pool skimmer to remove as much
feed as possible that was dropped into the containers by cows during the experimental period.
Containers were weighed and the weight was recorded. If the weight was less than
approximately 45 kg, the container was refilled in increments of 5 or 10 L, until the weight
reached greater than that number. Containers were then replaced in each feed bunk at the same
time and the behavior recording recommenced. Behavior recording ended 12 h after the
experimental period began. Containers were weighed, refilled and replaced in the bunks.
Initially, some cows pushed containers from the bunk in order to access feed. If
containers were pushed more than 20 cm from the bunk, they were pushed back into the bunk
within cow’s reach. If the cow routinely pushed the containers away, the containers were banded
to the bunks with rubber straps. Straps did not interfere with drinking. If the containers
contained less than approximately 10 L of water before the scheduled refilling time, containers
were removed from all bunks, and the containers were refilled so that containers none were ever
empty and cows were always able to drink from either container.
Containers were left in the bunks for cow’s to access for 22 h. They were checked at
0900 h, weighed and refilled. Containers were removed and weighed at the end of the
experimental period (1400 h).
Behavior Assessment. Drinking behaviors were recorded for the initial 12 h of the
experimental period and categorized for analyses. Categories were lapping, drinking, and ‘other’.
Drinking was defined as a cow fully or partially submerging her muzzle in water to consume
water. Lapping was defined as a cow consuming water without submerging her muzzle.

The

‘other’ category included all behaviors associated with containers when cows were not actively

150

consuming water. Non-consumption behavior interaction examples included splashing, sniffing
and pushing containers away. Frequency of interaction with water containers also was evaluated.
Behavior measurements were recorded to the nearest second. The beginning of one behavior
signified the end of the previous behavior, so only beginning time of behavior was recorded. If a
behavior was associated with a water container, the designated treatment identifier was recorded.
Other common behaviors recorded included eating, ruminating, urinating, defecating, standing
up, lying down, resting, and standing.

Preparation of Treatments
The 2% sodium chloride treatment was prepared by adding 2 g of sodium chloride to 100
ml of water. Sodium chloride was pre-measured into containers that held 200 g of NaCl each,
and were added to 10 L of water and mixed until NaCl dissolved into solution and then added to
the water container. These 10 L aliquots were mixed until the water container was filled.
Containers held approximately 50 L.

Water containers
Water was offered in two, 75 L capacity white Rubbermaid Brute™ (Rubbermaid
Commercial Products, Winchester, VA) storage tote containers (W: 38.4 cm, x H: 44.1 cm x L:
70.8 cm) placed in each the left and right sides of individual cow’s feed bunks with feed situated
between the containers. There was a vacant tie-stall between cows to avoid cow social
interactions that potentially might affect water intake.

151

Training
Cows were blocked from access to water in the milking parlor lane, so the only source of
water was the source provided in the tie-stall throughout the pre-experimental and experimental
periods. Nine d prior to the beginning of each experiment, cows were offered water from the onsite tap water source in the experimental water containers for 48 h to ensure that all cows became
accustomed to the containers. Containers were removed from the bunks, weighed and refilled
every 6 h during training. Every 24 h containers were removed, emptied, scrubbed with
detergent, rinsed well and refilled. There were no issues with cows refusing to drink from
containers. Following the initial training period cows were placed in tie-stalls with individual
automatic water cups, as in normal settings, and water intake was measured with in-line water
flow meters for 5 d to establish baseline water intake. Forty-eight h immediately prior to actual
data collection cows were moved to stalls without automatic water cups, and water was
accessible only in the experimental water containers to re-accustom the cows to the containers
and establish baseline intake from water containers. All cows were exposed to water containers
for the same length of time during pre-experimental acclimation periods and always located with
an unoccupied stall on each side.
Common Feeding, Water and Milk Data Collection
A basal diet formulated to meet or exceed NRC (2001) recommendations for all nutrients
was fed ad libitum throughout each experiment (Table 5.2). Cows were fed once daily in their
individual tie-stalls, and milked twice daily. Milk yield and water and feed intake were recorded
daily for the entire experimental period, including adjustment periods. Pre-experimental feed
intake, water intake, and milk yield and composition are presented in Table 5.1 for
characterization of experimental animals.
152

Feed and water intakes (in-line water flow meters) were measured daily throughout the
experiment. Individual feed ingredients were sampled the week before each collection period
began, and again between experimental periods. Each individual feed sample was weighed and
dried in a forced-air oven at 55°C for 72 h until completely dry to determine DM percentage.
Dried samples were ground through a 2 mm screen in a Wiley mill (Arthur H. Thomas,
Philadelphia, PA). Similar feed ingredient samples were pooled by experiment and sent for
analysis by Dairy One Laboratories (Ithaca, NY). Milk samples were taken from one AM and
PM milking and sent for composition analysis at Michigan DHIA Universal Laboratory Services
(Lansing, MI). Control water was sampled daily through the experiment and pooled into one
composite sample. Raw water sample was sent to Midwest Laboratories (Omaha, NE) for Direct
metals water analysis (EPA 200.7 method (1994)). A second sample was acidified (pH < 2.0)
with 1 mL nitric acid, and sent to the same laboratory for acid digestion and metals analyses
(Total recoverable metals analysis using EPA 200.7 method (1994)). Water analyses are
presented in Table 5.3. Internal water reference samples containing either 0, 1, 4 and 8 mg of
Fe/L from ferrous lactate were placed as unknowns within the set of samples in triplicate for
analysis. The intra-assay %CV was 4.0%.

Statistical Analyses.
Data were analyzed by method of least-squares ANOVA using the Mixed Models
procedure of SAS (Version 9.2, SAS Institute Inc.). Data were analyzed as completely
randomized design. Differences in water intake were compared between treatments. Data of

153

hourly measurements were analyzed using the REPEATED statement, and the covariance
structure that resulted in the lowest Bayesian information criterion (BIC) was used. Fixed effects
in the model included treatment, hour, and treatment by hour interaction. Random effects
included cow and period. If the interaction term was non-significant (P > 0.15), intake was
analyzed separately by hour. Residual distributions were tested for normality using the ShapiroWilk test, and outliers were identified using Cook’s D statistic. During evaluation of the full
model, including interaction terms, intake data were transformed by taking the square root to
accommodate the assumptions of normality. During analyses of individual hour water intake
responses and total behavior responses, if required, data also were transformed to square root,
log10 or natural log if required to accommodate the assumptions of normality. Results were back
transformed for presentation. Differences between treatments were identified using the PDIFF
option with Tukey’s adjustment. Least-squares means are presented. Significance for main
effects was declared at P < 0.05, and a trend toward significance at P < 0.10. Significance for
interaction effects was declared at P < 0.10, and a trend toward significance at P < 0.15.

154

RESULTS

There was a treatment by hour interaction on water intake (P <0.0001). This effect was
heavily influenced by intake responses in period 1, as cows greatly preferred NaCl treatment
water initially (Figure 5.1). When all periods were included in analyses, there was no difference
between Control and NaCl treatments after 2 h (P = 0.70). From h 4 through the end of the
experimental period, water intake of the Control treatment was greater than NaCl (P < 0.001;
Table 5.4). The drinking duration and drinking frequency of cows consuming Control water was
greater than cows consuming the NaCl treatment (P = 0.0006; P = 0.0007; Table 5.5). Cows
spent more total time lapping (P < 0.0001) and lapped more frequently (P < 0.0001) from the
NaCl treatment water than the Control water. Total duration (P = 0.91) and frequency of
behavioral interactions (P = 0.85) with containers, but not consuming water, was not different
between treatments.

155

DISCUSSION
We were initially concerned that cows would not choose between water treatments, but
instead immediately drink from the first container they encountered and not test both treatments.
For this reason we decided to offer cows treatments that had drastically different tastes; tap water
and 2% NaCl water. Additionally, cows respond to NaCl in water (Goatcher and Church, 1970b).
Cows initially preferred the NaCl treatment to the Control treatment in the first period
(Figure 1). After 15 min, cows already had consumed so much NaCl that the experimental period
had to be prematurely paused so that water containers could be refilled. This was unexpected as
Goatcher and Church (1970b) noted that cattle offered pairs of treatments with varying
concentrations of NaCl (0.2% to 1.25%) rejected water with NaCl concentrations greater than
0.32%. After the first 2 h of the initial period, cows exhibited the laxative effect characteristic of
NaCl toxicity (Jaster et al., 1978). They also began drinking large amounts of the Control water.
When all periods were included in the analyses water intakes at 2 h were not different between
treatments (Table 5.4). By h 6 Control water intake was greater than intake of NaCl treatment (P
= 0.0012). Intake of the Control treatment remained greater through the end of the experimental
period (P <0.0001).
For this reason, we decided that experimental periods should be extended beyond the 12
h period initially planned. Cow preference did not change beyond h 6. It also was not altered in
the two additional periods (Figure 2). We determined that experimental periods should be 22 h in
length, and 1 period (the initial period) was enough for cows to make a choice between two
treatments.

156

Cardot et al. (2008) determined that water intake was least during the evening hours. This
was a concern, as the initial 12 h of the experimental period took place from 1600 – 0400 h. We
were unsure if afternoon milking and a second feeding would be enough to induce water
consumption and allow us to find differences in treatment water intake. This was before we
decided to extend the experimental period beyond 12 h. The 2-h period necessary for preparation
of water containers and water treatments, and in combination with milking, was apparently a
long enough period without water intake to induce drinking.
We developed a procedure to manually measure drinking behavior. One of the difficulties
of recording data manually was that multiple behaviors took place within a short period of time,
which made recording accurate data difficult, if both a beginning and end time had to be
recorded. Recording continuously shortened the information that needed to be recorded. We
realized that it was easier and lead to more consistent data recording among observers to record
all behavior. Cows spent more time drinking Control water than NaCl water, and more time
lapping from the NaCl treatment water than Control water (Table 5.5). We were able to detect
treatment differences for drinking behavior duration, and frequencies as expected.

157

CONCLUSIONS
We determined that 22 h was adequate for cows to establish preference for a water
treatment. The 2-h period required to prepare water container and treatments in combination
with milking was long enough to drive thirst in concert with milking and feeding. Cows had
adequate thirst drive to establish preference through water intake, and cows tested both
treatments. Behavior was recorded manually continuously. Drinking behaviors were categorized
for analysis and differences between treatments for lapping and drinking durations and
frequencies were detected successfully.

158

TABLES
Table 5.1. Feed dry matter intake, milk yield, and water intake of cows during the preexperimental period
Mean
DMI, kg DM

±SD

1

23

1.3

Water intake, L

114.9

9.93

Milk yield, kg

52.8

6.72

1

SD = standard deviation.

159

Table 5.2. Ingredient and nutrient composition of diets fed
Ingredient
--------------Percent of dietary DM---------Chopped alfalfa hay
3.72
Corn silage
35.32
Whole cottonseed
5.95
Energy booster
0.40
Grass hay
5.02
Ground corn
9.53
Alfalfa haylage
9.51
High moisture corn
7.14
Soybean meal, 48% CP
14.72
Soyhulls
5.35
1
Mineral-vitamin mix
3.34
Nutrient Composition
DM, %
49.0
-------------Percent of dietary DM----------Neutral detergent fiber
33.9
Acid detergent fiber
22.4
Crude protein
17.58
Ca
1.01
P
0.37
K
1.38
Na
0.34
S
0.24
Mg
0.25
-------------------mg/kg-------------------Co
0.17
Cu
10.2
Fe
253
Mo
2.3
Mn
44
Zn
41
1

Mineral-vitamin mix contained 47.5% limestone, 22.5% sodium bicarbonate, 10.1% urea

(45%N), 8.2% magnesium sulfate, 7.5% sodium chloride, 1.6% biotin premix (1.4 g/kg), 1.3%
trace mineral premix, 30 KIU/kg vitamin A, 8 KIU/kg vitamin D, 56 KIU/kg vitamin E and 0.3%
Se-yeast.

160

Table 5.3. Water quality constituent analyses of Control drinking water
Control
Quality Constituent
TDS

Caution level

1

--------------------mg/L--------------------

2

460

1,000

Ca

114

150

Cl

24

500

Cu

ND

Fe (direct)

4

3

0.3

0.05

0.3

0.28

-

35.3

80

Mn

0.22

-

NO3-N

ND

Fe (total recoverable)

5

Mg
5

25

Na

9.3

150

SO4

97

300

Zn

0.31

-

Conductivity, mmhos/cm

0.71

1.5

pH

7.7

6.5-9

1
2
3
4

5

Caution levels from Midwest Laboratories (Omaha, NE)
TDS = total dissolved solids
ND = not detected
Direct metals analysis of raw (without acid digestion) water by Livestock Water Analysis using

EPA 200.7 method (1994) by Midwest Laboratories (Omaha, NE).
5

Total recoverable metals analysis of acidified water after acid digestion using EPA 200.7

method (1994) by Midwest Laboratories (Omaha, NE).

161

Table 5.4. Cumulative water intake (L) of cows pooled for all 3 periods
Treatments

1
2

Control

NaCl

SEM

P-value

2

6.8

5.3

1.79

0.70

6

52.3

26.9

5.07

0.0012

12

85.4

35.0

5.69

<0.0001

17

102.1

39.2

7.55

<0.0001

22

136.6

45.0

9.75

<0.0001

Hour
3

1
2
3

Treatments: Control = tap water; NaCl = 2% sodium chloride.
SEM = standard error of the mean.
Hour 2 data were natural log transformed for analysis, and LS means and SEM were back-

transformed for presentation in this table.
Table 5.5. Cumulative drinking behavior durations and frequencies of cows
Treatments

1
2

Control

NaCl

SEM

P-value

693

202

85.0

0.0006

295

1292

142.2

<0.0001

3

3

1.1

0.91

Drinking

18.4

6.2

3.20

0.0007

Lapping

10.6

28.1

4.48

<0.0001

5.8

5.3

1.90

0.85

Behavior Duration (s)
Drinking

3

Lapping
Other

4,5

Behavior Frequency

Other
1
2
3

4

Treatments: Control = tap water; NaCl = 2% sodium chloride.
SEM = standard error of the mean.
Data were log10 transformed for analysis and LS means and SEM were back-transformed for

presentation in this table.
4
5

Other = interacting with water container but not consuming water.
Data were square root transformed for analysis and LS means and SEM were back-transformed

for presentation in this table.
162

FIGURES

1

2

Figure 5.1. Treatment by hour (SEM = 8.03; P < 0.0001) response of water intake in period 1
140

Water Intake (L)

120
100
80
Control

60

NaCl

40
20
0
2

4

6

8

10

12

Hour
1
2

Treatments: Control = tap water; NaCl = 2% sodium chloride.
SEM = standard error of the mean.

163

17

22

1

2

Figure 5.2. Treatment by hour (SEM = 6.68; P < 0.0001) response of water intake across all
periods
160
140
Water Intake (L)

120
100
80
Control

60

NaCl

40
20
0
2

4

6

8

10

12

Hour
1
2

Treatments: Control = tap water; NaCl = 2% sodium chloride.
SEM = standard error of the mean.

164

17

22

REFERENCES

165

REFERENCES
Cardot, V., Y. Le Roux, and S. Jurjanz. 2008. Drinking behavior of lactating dairy cows and
prediction of their water intake. J. Dairy Sci. 91:2257-2264.
Coppock, E., R. W. Everett, N. E. Smith, S. T. Slack, and J. P. Harner. 1974. Variation in forage
preference in dairy cattle. J. Anim. Sci. 39:1170-1179.
Goatcher, W. D., and D. C. Church. 1970a. Taste responses in ruminants. II. Reactions of sheep
to acids, quinine, urea and sodium hydroxide. 30:784-790.
Goatcher, W. D., and D. C. Church. 1970b. Taste responses in ruminants. III. Reactions of
pygmy goats, normal goats, sheep and cattle to sucrose and sodium chloride. J. Anim.
Sci. 31:364-372.
Jaster, E. H., J. D. Schuh, and T. N. Wegner. 1978. Physiological effects of saline drinking water
on high producing dairy cows. J. Dairy Sci. 61:66-71.
National Research Council. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. Natl. Acad.
Sci., Washington, DC.

166