The magnesium allowance of the lactating cow


Material Information

The magnesium allowance of the lactating cow renal physiology and gastrointestinal tract absorption of magnesium as influenced by mineral ions and ionophores
Physical Description:
vii, 200 leaves : ill. ; 28 cm.
O'Connor, Anita Marie, 1954-
Publication Date:


Subjects / Keywords:
Dairy cattle -- Feeding and feeds   ( lcsh )
Minerals in animal nutrition   ( lcsh )
Magnesium in animal nutrition   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1987.
Includes bibliographical references (leaves 184-199).
Statement of Responsibility:
by Anita Marie O'Connor.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000951398
notis - AER3641
oclc - 16956320
System ID:

Full Text








Sincere appreciation and thanks are extended to David K. Beede,

Associate Professor of Dairy Science, for his untiring advice,

assistance, and guidance throughout the course of these studies. The

author is grateful to Dr. Charles J. Wilcox for invaluable assistance in

designing experiments and statistical analyses. Deep appreciation is

extended to Dr. Bruce R. Stevens for timely suggestions and advice

expressed throughout the course of the brush border membrane vesicle

experiments. Special recognition is expressed to committee members Dr.

Kimon J. Angelides, Dr. Clarence B. Ammerman, and Dr. Douglas B. Bates

for helpful comments, suggestions, and use of laboratory facilities.

The author appreciates the timely advice and encouragement of Dr.

Herbert Head and Dr. Roger P. Natzke throughout her graduate program.

The author is indebted to the farm crew at the Dairy Research Unit

in Hague for cooperation and assistance in large animal experiments.

Special thanks go to Dale Hissem and Austin Green. Appreciation is

extended to fellow graduate students and support staff for technical and

secretarial assistance. The author acknowledges gratefully the help of

Fran Romero, Lokenga Badinga, Jody Maley, Abu Bakar Chik, Changzheng

Wang, Estelle Hirchert, Paula Dolder, Glenda Walton, Mary Ellen Hissem,

and Susan Allen.

Appreciation is extended to Boliden Kemi AB for partial financial

support of the lactation study.



ACKNOWLEDGMENTS.. .............. .....................................ii

ABSTRACT ......... .................................


1 INTRODUCTION.................................................1

2 REVIEW OF THE LITERATURE......................................4

Role of Magnesium in the Ruminant.............................4
Importance of Magnesium ....................... ...............4
Methods of Determining Bioavailability......................8
Factors Affecting Bioavailability.......................... 11
Bioavailability in Supplements...........................12
Bioavailability in Forages and
Utilization of Magnesium by the Ruminant................... 17
Site and Mechanism of Absorption.........................17
Relationships of Sodium and Potassium to
Magnesium Absorption and Turnover......................20
Metabolism in Tissues.....................................21
Metabolism: Excretion Patterns.............................22
The Lactating Cow and Magnesium..............................24
In Vitro Methods of Determining Magnesium
Absorption by the Gut......................................26
Ussing Chamber Approach..................................26
Brush Border Membrane Vesicles..............................27
lonophores and the Ruminant.................................30
Mode of Action of lonophores in the Whole
Animal............ .................. ......... ..........31
Effects of lonophores on Mineral Metabolism
and Absorption in the Ruminant..........................33
Mode of Action at the Cellular Level.....................35

THE RUMENS OF LACTATING DAIRY COWS...........................40

Materials and Methods.......................................41
Animals and Design........................................41


Sampling and Analyses......................................45
Calculations................................. ............ 48
Statistical Analyses....................................48
Results and Discussion....................................48
Summary.................................. ..................56

DAIRY COWS.............................. ....... ........... 58

Materials and Methods......................................60
Experimental Design and Treatments.........................60
Animals and Diets..........................................60
Sampling and Analyses...................... .............60
Calculations.................. ........................61
Statistical Analyses.....................................o2
Results and Discussion.......................................62
Summary....... .......... ...................... ............71


Introduction............................ ........... 72
Materials and Methods..............................74
Results and Discussion.......................................79
Intake, Milk Yield and Composition Responses...............79
Milk and Plasma Mineral Concentrations and
Fractional Excretions.....................................85

IN USSING CHAMBERS.........................................99

Introduction........... .... .................................... 99
Materials and Methods........................................00
Results and Discussion...................................103
Summary....................................... ............ .. 112

MEMBRANE VESICLES...........................................114

Materials and Methods........................................115
Membrane Preparations....................................115
Transport Experiments.....................................116
Statistical Analyses....................................... 118

Results and Discussion.....................................119
Glucose Transport Experiments............................. 119
Calcium Transport Experiments.............................119
Magnesium Transport Experiments...........................131
Michaelis-Menten Kinetics................................139
Summary ......................... .................... .. 145

8 GENERAL DISCUSSION AND SUMMARY..............................146

Lactational and Dry Matter Intake Responses
to Minerals.............................................146
Metabolic Responses to Magnesium, Potassium
and Sodium...............................................148
Ionophore Effects on Calcium and Magnesium
Uptake at the Brush Border...............................150
Future Research..............................................152


A GENERAL LINEAR MODELS FOR CHAPTER 3.........................155
B GENERAL LINEAR MODESL FOR CHAPTER 4.........................163
C GENERAL LINEAR MODELS FOR CHAPTER 5.........................175
FOR CHAPTER 7............................... .............177

REFERENCES........................................................ 184

BIOGRAPHICAL SKETCH...............................................200

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Anita Marie O'Connor

May, 1987

Chairman: David K. Beede
Major Department: Animal Science

Two experiments with lactating cows were conducted to evaluate

effects of higher dietary magnesium (Mg), potassium (K), and sodium (Na)

on ruminal absorption and renal processing of Mg and other

macroelements, and on lactational performance. In vitro experiments

using jejeunal brush border membrane vesicles and duodenal and ruminal

tissue sections in Ussing chambers were done to study effects of

ionophores on absorption of Mg, Ca, and other macroelements.

In experiment 1 eight lactating cows were administered .20 or .40%

Mg, .87 or 1.65% K and .22 or .60% Na in dietary dry matter. Higher

dietary Na and Mg interacted to increase ruminal disappearance

(absorption) of Mg. Higher K decreased the molar Na:K ratio in the

rumen. There were no main effects of treatments on ruminal liquid

dilution rates or outflows; however, higher Mg and K interacted to

increase daily ruminal liquid outflows (two-factor interaction). Higher

K tended to increase net tubular reabsorption of Mg and decrease

fractional excretion of Mg. Higher Mg and K interacted to decrease

daily excretion and clearance of Mg (two-factor interactions).

Experiment 2 was designed to study the Mg allowance of the lactating

cow, and two- and three-factor interactions with Na and K. Forty-eight

mid-lactation cows were fed .26, .38, .48 or .60% Mg, .24 or .62% Na and

1.14 or 1.59% K in a 4 x 2 x 2 factorial arrangement. Actual and

fat-corrected milk yields were increased as dietary Mg increased from

.26 to .48% Mg; there were no effects of Mg on milk fat percent. Feed

intakes were equal with .26, .38 and .48% Mg but declined with .60% Mg.

There were no effects of higher Na or K on feed intake or milk

production. Fractional excretion of Ca declined with higher K and

increased with higher Na.

Effects of lasalocid and monensin on apparent absorption of Mg, K,

Na, calcium (Ca), and phosphorus (P) were evaluated in vitro with.

duodenal and ruminal tissue sections mounted in Ussing chambers in the

third set of experiments. There were trends toward increased mineral

absorption due to ionophore treatments. However, due to technical

problems with the experiments the data were inconclusive.

In the fourth set of experiments, effects of ionophores A23187,

lasalocid, lysocellin and monensin on Mg and Ca uptake by brush border

membrane vesicles were studied. The putative Ca ionophore A23187

increased Ca and Mg uptake by brush border membrane vesicles. There

were trends towards increased Ca and Mg uptake by brush border membrane

vesicles due to lasalocid, and no effects due to either monensin or



Early interest in magnesium (Mg) was based on the observation that

the ion was linked to hypomagnesemic tetany. Most Mg research with

ruminants to date has focused on methods of eliminating this disease.

Limited data exist on the Mg requirement per se for the lactating cow.

Requirements set by the National Research Council (NRC, 1978) and

Agriculture Research Council (ARC, 1980) were based on early experiments

with low producing lactating cows (Blaxter and McGill, 1956; Rook and

Balch, 1958; Rook et al.,1958; Van der Horst, 1960; Kemp et al., 1961).

In this country the genetic capacity of dairy cattle to produce milk has

improved tremendously, largely through artificial insemination and

superior techniques to identify top bulls (Warwick and Legates, 1979).

As cattle continue to improve, it is important to continuously evaluate

nutritional requirements. Thus, it is necessary to study the Mg

requirements of the superior cows we have today for producing milk.

It is of equal importance to determine both the requirements and

effects of dietary factors on absorption of Mg in the lactating cow.

These factors may include other electrolytes such as sodium (Na) or

potassium (K), or feed additives such as the ionophores.

In the past 10 yr Na has increased in diets fed to lactating cows

largely through inclusion of sodium bicarbonate (NaHCO3) as a buffer.

Sodium may potentially affect the absorption and metabolism of other

minerals in lactating cows. Sodium increased blood buffering capacity

and depressed plasma and milk concentrations of Na (Schneider et al.,

1986). The molar Na:K ratio in the rumen was reported to be important

in promoting ruminal absorption of Mg (Martens and Rayssiguier, 1980).

Thus, Na supplied by NaHCO3 may aid in enhancing ruminal absorption of


Effects of K on Mg absorption and metabolism are well documented in

contrast to Na (Tomas and Potter, 1976a; Greene et al., 1983).

Hypomagnesemic tetany was associated with animals grazing pastures high

in K content, and several studies demonstrated that excessive dietary K

intake decreased Mg absorption of Mg (Suttle and Field, 1969; Fontenot

et al., 1973). In the ruminant, the kidney minimized Mg excretion when

diets with high K content were fed, or K was infused intravenously

(Deetz et al., 1981). There is increasing evidence that higher dietary

K may be required by heat-stressed animals. Also, additional K

supplementation may be necessary for animals fed by-product feeds with

low K content (Schneider et al., 1984, 1986). Thus, it is important to

determine effects these trends in feeding practices have on Mg balance

and feeding recommendations.

In recent years the ionophores received considerable attention as

feed additives. These compounds are known to alter ruminal

fermentation, increase metabolic efficiency and may decrease ruminal

protein degradation (Goodrich et al., 1984; Schelling, 1984). Limited

balance studies with ruminants suggested that specific ionophores

enhanced calcium (Ca), K, Mg and(or) Na absorption (Donoho et al., 1978;

Starnes et al., 1984). Ionophores primarily have been used in the beef

cattle industry. However, they may become important to the dairy

industry, as new methods for feeding high producing cows are explored.

They may enhance gut absorption of Ca and(or) Mg or other nutrients

critical to milk production. To date there is virtually no information

available on effects of ionophores on nutrient absorption by lactating


Purpose of the experiments reported herein was to gain basic

information on the absorption and metabolism of Mg in the lactating cow.

Specific objectives were

(1) to study effects of elevated dietary Na, K, and Mg on apparent

ruminal disappearance of Mg;

(2) to evaluate effects of elevated dietary Na, K, and Mg on renal

excretion and metabolism of Mg;

(3) to determine effects of higher dietary Na, K, and Mg on

lactational performance; and

(4) to study effects of several ionophores on mineral absorption by

ruminal and duodenal tissue in vitro with Ussing chambers, and

at the membrane level with bovine jejeunal brush border

membrane vesicles.


Role of Magnesium in the Ruminant

The first report associating magnesium (Mg) with grass tetany or

hypomagnesemia was published in 1929 (Sjollema and Seekles, 1929, cited

through Kemp, 1983). Since this disease is of economic importance in

many parts of the world there has been a plethora of experiments to

determine causs(s) of grass tetany. Research focused on.relationships

of plant and soil Mg, and Mg metabolism in ruminants. For the most part

these studies utilized sheep and cattle with fewer experiments with

lactating dairy cows.

Importance of Magnesium

Magnesium is found ubiquitously in domestic animals; the only

cations having higher total body quantities than Mg are sodium (Na),

(Ca), and potassium (K). It is similar to K in that it is predominantly

located intracellularly. Approximately 60 to 70% of total body Mg is

found in the skeletal system where it is adsorbed, in part, onto

hydroxyapatite crystal. Remaining Mg (30%) is located in the soft

tissues and circulating body fluids. Blood Mg is found in highest

concentration in red cells, with intermediate amounts in serum and less

bound to albumin and globulin, respectively.

Magnesium is important biochemically because it serves as a cofactor

for numerous enzymes and is required for a plethora of reactions in the

mainstream of carbohydrate, lipid, and protein metabolism. Most notably

it is a cofactor for adenosine triphosphatase (ATPase) (Lehninger,

1970). It is required also by enzymes associated with neurotransmitter

release, deoxyribonucleic acid (DNA) and ribonucleic acid (RNA)

synthesis, and methyl group transfer. Magnesium is critical to protein

and carbohydrate metabolism. Magnesium-deficient rats exhibited

hypoproteinemia, hypogammaglobulinemia, and increased plasma urea

(Rayssiguier et al., 1977, and Larvor and Labat, 1978, cited through

Larvor, 1983). In vitro amino acid uptake by tissues from Mg-deficient

rats significantly decreased (Amarial and Nurti, 1972, cited through

Larvor, 1983). Impairment in protein metabolism may occur because

several factors needed for initiation and elongation of the nascent

polypeptide chain on the ribosome require Mg (Larvor, 1983). In the

Mg-deficient calf decreased glucose-6-phosphate dehydrogenase activity

decreased red blood cell glycolysis causing subsequent anemia (Mishra

and Perey, 1960, cited through Larvor, 1983).

The importance of Mg physiologically is evident by the number of

disorders which occurred in Mg-deficient animals. These included

impaired growth, retarded immune and allergic ability, muscular

contraction, red blood cell integrity, neoplasms, impaired metabolism of

collagen-rich tissues, kidney calcification, and disturbances in calcium

metabolism (Larvor, 1983). Most of these disorders were demonstrated in

the rat. Exceptions are derangements in muscle contraction, red blood

cell stability and disorders in Ca metabolism, which were shown in the

ruminant (Larvor, 1983). Increased and sustained muscle contractions,

typically seen in grass tetany, possibly were caused by leakage of Ca

from the Mg requiring Ca pump in the sarcoplasmic reticulum in skeletal

muscle (Larvor, 1983). In Mg-deficient sheep amplitude and frequency of

contractions of the bladder, cecum and colon were reduced, and amplitude

of contractions in the small intestine were reduced (Larvor, 1983).

Anemia in the Ca-deficient calf may be caused by decreased stability of

red blood cells due to decreased activity of glucose-6-phosphate

dehydrogenase (Larvor, 1983). The Mg-deficient calf was reported to

become hypocalcemic; in these animals total Ca pool was unchanged with

diminished exchange of plasma Ca with labile osseus tissue Ca (Larvor,

1983). Also, the skeletal system appeared less sensitive to parathyroid

hormone and Vitamin D3; this may have occurred due to impaired

collagen synthesis in the tissues which are dependent on several

Mg-requiring enzymes (Larvor, 1983).

Magnesium was important in maintaining adequate dry matter intake in

the ruminant. Magnesium supplementation improved dry matter intake in

animals on both Mg-deficient and sufficient diets (Ammerman et al.,

1971; Chicco et al., 1973a; Reid, 1979,1983; Vona et al., 1980).

Experiments demonstrated that dry matter intake was reduced 76% in lambs

and 58% in adult sheep within 3 to 4 d of feeding a Mg-deficient diet

(Ammerman et al., 1971; Chicco et al., 1973a); restoration of dietary Mg

immediately improved dry matter consumption. Pregnant beef cows fed

corn stover diets and .45% (dry basis) Mg from magnesium sulfate

(MgSO4) had higher serum Mg and increased body weight gains compared

to cows supplemented with .16% Mg (dry basis) (Vona et al., 1980). In

dry beef cows, MgSO significantly increased stover intake and serum

Mg (Vona et al., 1980). The intake response to Mg was noted in sheep

fed timothy hay supplemented with MgSO (Reid, 1983). Generally,

magnesium oxide (MgO) has not been efficacious in increasing dry matter

intake (Thomas and Emery, 1969b; Stout and Bush, 1972; Thomas et al.

1984). Magnesium oxide reduced dry matter intake in dairy cattle,

possibly due to excessive dietary Mg (.6 to .7%, dry basis) (Thomas et

al. 1984). The question of why and how Mg elicits the feeding response

remains an enigma. Inappetance with Mg deficient diets may be partly

related to a high requirement by ruminal microorganisms for Mg.

Cellulolytic digestion in vitro and in vivo decreased in the absence of

Mg (Ammerman et al., 1971). Cellulytic digestion in washed suspensions

of ruminal microorganisms improved moderately when Mg concentration in

the media was increased (Ammerman et al., 1971). Chamberlain and

Burroughs (1962) and Wilson (1980) showed supplemental MgSO4 addition

to incubation mixtures improved ruminal fermentation. However, intake

responses cannot be explained entirely by satisfying microbial

requirements for Mg since they have been noted in animals having high Mg

intakes well in excess of suggested minimum daily requirements. In

experiments in which intake of alfalfa was improved by fertilization

with kiersite (MgSO ), dietary concentrations of Mg were higher

generally than .3% of dry matter (Reid et al., 1979). Improvement of

dry matter intakes by Mg may be due, in part, to an effect at the tissue

level rather than simply enhanced diet palatability or improved

fermentation (Allsop and Rook, 1979). Hypomagnesemic sheep only resumed

eating after an intravenous infusion of magnesium chloride (MgC 1)

(Allsop and Rook, 1979). The feeding response to Mg may be due to

increased concentration of Mg in cerebral spinal fluid (Seoane and

Baile, 1973). Injecting 20 umol of Ca or Mg into the lateral ventricles

of sheep elicited a threefold to fourfold increase in dry matter

consumption within 2 h. Hypomagnesemia seen in with grass tetany was

believed to be related to low cerebral fluid Mg (Littledike et al.,


Methods of Determining Bioavailability

Several methods exist for measurement of bioavailability of Mg.

These include in situ, true absorption, and apparent absorption

techniques, and analysis of plasma and urinary Mg excretion patterns.

In the in situ technique, samples of Mg supplements were suspended

in small synthetic bags in the rumens of cattle; bags withdrawn at

specific times were analyzed for Mg content (Wilson, 1981). A time

course for Mg disappearance within the bags was plotted to determine

first order rates of disappearance (Wilson, 1981). For purposes of

ranking Mg supplements the situ technique is simple and rapid. Wilson

(1981) incubated 27 MgO supplements in situ and found 48 h solubilities

were highly correlated to bioavailability (e.g., apparent absorption

measured by three different methods). Rates of disappearance were

lower when cattle were fed "indoor" rations versus pasture grass,

perhaps reflecting greater solubility of supplements in cattle with

higher ruminal liquid volumes (Wilson, 1981). A major disadvantage of

the in situ technique is that it cannot evaluate powdered mineral

sources materials due to rapid loss of the supplement through the bag

pores (Wilson, 1981).

Apparent absorption techniques usually involve the total collection

(balance method) approach to measurement of bioavailability. This

requires determination of Mg content of dry matter intake and feces, and

apparent absorption is calculated as the difference between Mg content

of feces and the diet. Apparent absorption frequently has been used to

determine bioavailability of Mg; yet several problems exist with this

technique. Foremost is the uncertainty of contribution of endogenous Mg

and consequent inaccuracies of apparent absorption estimates. Secondly,

there may be analytical errors in dietary inputs and excreta of Mg which

can lead to erroneously high mineral retention values (Fishwick and

Hemmingway, 1973) because mineral retention in the balance method is

calculated by difference and necessarily involves cumulative errors

(Wilson, 1981). Recent work has indicated that use of chromic oxide as

an external marker to calculate apparent absorption from mineral and

chromium analysis of fecal grab samples is an easy and accurate method

of determining bioavailability (Wilson, 1981). This is a particularly

useful technique for experiments involving large numbers of animals

under field conditions (Wilson, 1981).

To determine amount of Mg truely absorbed the researcher must know

the contribution of endogenous Mg and there are a variety of methods

available to do this. These estimates are more accurate than apparent

absorption and are generally higher (Chicco et al., 1972). Calculation

of true absorption of Mg in ruminants is limited because of problems in

estimating endogenous Mg in the gut, which can be sizable and highly

variable. One approach for measurement of endogenous Mg is the isotope

dilution method using 28Mg (Care, 1960; Field, 1961b; McAleese et al.,

1961; Care, 1965; House and Van Campen, 1971; Larvor, 1976) in which

radioactive Mg equilibrates within the animal and endogenous Mg is

calculated from dilution of the isotope in the feces. Experiments with
28Mg are nil in dairy ca e due to its short half-life (21 h) and
Mg are nil in dairy cattle due to its short half-life (21 h) and

considerable expense in utilizing the isotope in large animals. Other

methods for estimation of endogenous Mg secretion into the gut involve

feeding a series of purified or conventional diets of differing Mg

content and regressing fecal Mg on Mg intake. Endogenous Mg is found by

extrapolation to zero Mg intake (Chicco et al, 1972; Allsop and Rook,

1979). Problems with this approach include low palatability of purified

diets, errors involved in estimating factors used in regression

equations, and conservation of endogenous Mg during deprivation

(McAleese et al., 1961; House and Van Campen, 1971).

Estimating the bioavailability of Mg by evaluation of plasma

involves taking sequential samples and estimating extent to which plasma

Mg is elevated by supplementation (Wilson, 1981). Measurement of plasma

Mg concentration, as an index of bioavailability, is not the most

sensitive method of estimating Mg absorption and is effective only when

ruminants are fed Mg-deficient diets (Gentry et al., 1978; Wilson,

1981). A problem with using plasma Mg to determine bioavailability is

that a low Mg diet must be formulated and fed in order to make animals

hypomagnesemic before testing the supplements; this can be

time-consuming and laborious. Wilson (1981) noted a high correlation

between plasma estimates of bioavailability of Mg and values obtained

using the chromic oxide approach for a series of MgO supplements.

Plasma Mg commonly is used as an index of Mg status (Grace and MacRae,

1972; Playne et al., 1978; Wilson and Grace, 1978).

Measurement of urinary Mg excretion is a sensitive technique for

measurement of both Mg status and relative efficiency of absorption

since bioavailability of supplements can be ranked on the basis of their

ability to increase Mg excretion (Field et al., 1958; Field, 1962;

Storry and Rook, 1963; Kemp and Geurink, 1978; Wilson, 1981; Field,

1983; Kemp, 1983). Relative efficiency of absorption of several Mg

supplements can be found by comparing regressions of Mg intake on renal

Mg excretion rather than absolute quantities (Field et al., 1958).

Excess dietary Mg is excreted readily when the renal threshold of

approximately 2 mg/100 ml is exceeded,resulting in a linear relationship

between apparent absorption of Mg and urinary excretion (Rook et al.,

1958). Jesse et al. (1981) found MgO sources of greatest availability

to be excreted within 24 h of oral dosing whereas slowly available

sources took 3 to 4 d to be excreted in the urine.

In summary, several methods exist for determination of

bioavailability of Mg supplements. Wilson (1981) found that

bioavailabilities when measured by the chromic oxide technique were in

agreement with those found using plasma concentrations of minerals. In

situ bioavailabilities also agreed with results of trials using chromic

oxide as a marker. Thus, use of in situ, plasma mineral concentration

values and the chromic oxide technique may give reasonable estimates of

Mg bioavailabilities. However, it should be considered that the

comparisons made in these experiments were not performed with the same

group of animals and are therefore perhaps not as accurate as they could


Factors Affecting Bioavailability

A variety of factors have been suggested to affect bioavailability

of Mg: diet composition, Mg status of the animal, ruminal pH, solubility

of Mg in ruminal fluid, Mg distribution in the solid ruminal phase,

buffering effects in the rumen, Ca, phosphorus (P), Na, K, manganese

(Mn), citrate, phytic acid, chelators, changes in feeding systems, age

and genetic differences of animals, vitamin D3, nitrogen (N), water

intake, transmural ruminal potential, and ruminal volatile fatty acid

(VFA) concentrations (Peeler, 1972; Field, 1983). Additionally, there

are several physical factors of Mg supplements which can affect


Bioavailability in supplements

Bioavailability of Mg supplements may be dependent on particle size,

solubility of the supplement, ruminal pH, temperature of calcination,

geological sources, and hardness or color (Wilson, 1981).

Smaller particle size may render Mg supplements more available to

the animal. Jesse et al. (1981) noted that MgO supplements having

smaller particle sizes (850 um) were more readily excreted in the urine

than larger sizes which took 3 to 4 d to be absorbed. The smallest

particle size tested (850 um) increased in vitro ruminal pH and

concentration of Mg, and significantly increased in vivo ruminal acetate

concentration while decreasing propionate. Wilson (1981) showed

bioavailability (e.g., apparent absorption measured by balance, chromic

oxide technique and plasma Mg) was always higher for sources of MgO with

smaller particle size, in both sheep and cattle under a variety of

feeding regimens. Average bioavailability of three Spanish MgO sources

was 33.8% for the less than 75 um size, and 27.8% and 15.0% for the 150

to 250 um and 500 to 1000 um sizes, respectively (Wilson, 1981). In

comparing MgO supplements from different countries, mean bioavailability

for the powdered MgO supplements (less than 75 um) was 38.8% compared

with 17.7% for the granular form (greater than 75 um) (Wilson, 1981).

Magnesium oxides from different countries (sources) had different

bioavailabilities within particle sizes (Wilson, 1981). Thus, although

particle size is an important determinant of bioavailability of Mg

supplements, source also must be considered (Wilson, 1981). Smaller

particle sizes of MgO significantly increased Mg concentration in

ruminal fluid and urinary Mg (Thomas et al., 1984). Henry et al.(1986)

recently reported that smaller particle size of Mg supplements promotes

greater bioavailability as determined in balance trials with wethers.

Greater solubility of the Mg supplement may indicate greater

bioavailability to the animal. Several studies found that smaller

particle size of MgO was associated with greater solubility of Mg in

ruminal fluid in vitro and in vivo (Jesse et al.,1981). Solubility in

ruminal fluid generally is associated with higher urinary Mg

concentrations indicating greater absorption (Jesse et al., 1981).

Wilson (1981) also found a positive relationship between solubility of

supplements, defined as release from nylon bags suspended

intraruminally,and bioavailability, when comparing five different MgO


There is a negative relationship between ruminal pH and solubility

of Mg in rumen fluid; above pH 6.0 ultrafiltrability of Mg rapidly

decreases (Smith and Horn, 1976). When varying amounts of sulfuric acid

were added to rumen contents of steers in vitro, ultrafiltrable Mg

increased with acid addition. Greatest ultrafiltrability of Mg occurred

below pH 6.6 to 7.0, above this point Smith and Horn (1976) speculated

that Mg may precipitate as magnesium ammonium phosphate or mixed Mg or

Ca phosphates or both (Smith and Horn, 1976). Storry (1961b)

demonstrated that lower pH in any location in the gastrointestinal tract

was positively correlated with ultrafiltrability of Mg. Interest in the

positive relationship between ruminal fluid pH and Mg solubility is due

to the large body of evidence which showed greater absorption occurred

with a highly soluble Mg source (Field, 1983). However, in an

experiment with steers, when .3 M sulfuric acid was infused before and

after feeding, ruminal pH dropped approximately .6 units and no

differences in 28Mg absorption were seen. Authors attributed this to

small numbers of animals used, great variation between animals and the

inherently high solubility of the Mg source used in the experiment

(Smith and Horn, 1976).

Temperature of calcination plays an important role in

bioavailability of MgO sources. Wilson (1981) compared raw magnesite

and magnesites calcined for .75 h at 500, 650, 800, 900 and 1100 C in

sheep fed grass nuts or concentrate cubes. Bioavailabilities (e.g.,

calculated by the balance method) were less than 20% for raw, 500, and

650 C calcined MgO sources, but ranged from 45 to 47% for 800, 900 and

1100 C MgO supplements in animals fed concentrate cubes and 28 to 38% in

animals fed grass nuts. Treatment of MgO at 1,300 C for 3 h rendered

the product approximately 30% bioavailable. Thus, optimum calcination

temperature, in terms of bioavailability, appeared to be within the

range of 800 to 1100 C with limited bioavailability for supplements

calcined above 1100 C (Wilson, 1981).

Differences in the bioavailabilities of Mg supplements may occur

because of substantial variation within product type. In the case of

MgO there is extensive variation in geological source, hardness, color,

temperature of calcination and particle size (Hemmingway, 1985). There

is greater variation in the MgO products in contrast to MgP.

Summarizing multiple experiments with MgP, Hemmingway (1985) concluded

mean bioavailability of feed grade MgP was 34.9% + 5.2% for four MgP

sources versus 24.7 + 11.3% for 20 commercial preparations of MgO

(Hemmingway, 1985). Magnesium phosphate is an Mg source which has met

with success both as a Mg and P source in Europe (Fishwick and

Hemmingway, 1973; Richie and Fishwick, 1977; Fishwick, 1978; Hemmingway

and McLaughlin, 1979; Wilson, 1981).

Bioavailability in forages and concentrates

Magnesium content in forages ranged from very low to .35% of dry

matter (Rook et al., 1958; Kemp et al. 1961; Field, 1962; Rook and

Campling 1962; Rook et al., 1964; Joyce and Rattray, 1970; Grace et al.,

1974; Gross and Jung, 1978). Accumulation of forage Mg depends on a

variety of factors including environmental temperature, soil conditions,

forage species, forage maturity and fertilization with Mg, K, and N

(Reid, 1983). Bioavailability of Mg in forages is lower than in

concentrates and supplements, and highly variable among plant species.

It is not correlated highly with Mg concentration in plants (Reid,


Kemp and Geurink (1978) estimated bioavailability of Mg in typical

lactating cow diets ranged between 7 and 33% with a mean of 17%.

Bioavailability of Mg was lower than P (27%) and Ca (30%) and

substantially lower than sulfur (S), chloride (Cl), Na, and K which

ranged between 73 and 89% (Kemp and Geurink, 1978). Bioavailability was

correlated positively with Ca absorption (Lomba et al., 1968).

Magnesium absorption increased with forage maturity in contrast to Ca,

P, and S (Powell et al., 1978).

Magnesium is digested at a similar rate as S, P, Ca, Na and K in the

rumen (Playne et al., 1978). This is perplexing in view of the fact

that Mg bioavailability in forages is minimal. Diminished

bioavailability of Mg was not related to chlorophyll content (Todd,

1961). Greater proportion of Mg than previously realized may be released

from forages but is unavailable due to adsorption onto either bacterial

cell walls or the solid phase of ruminal ingest (e.g., plant tissue,

Fitt et al., i972; 1974).

In efforts to improve plant Mg content and ultimately absorption and

net retention, kiersite was applied successfully to a variety of grasses

and legumes grown in Pennsylvania and West Virginia (Reid et al 1975;

Reid et al., 1978; Reid, 1979; Reid et al., 1979). Magnesium

fertilization generally increased concentration of Mg in forages but was

highly variable and inconsistent within forage species (Reid et al.

1975; Reid et al. 1978; Reid, 1979). Treatment with kiersite had no

effect on Ca, P, K or S balance in sheep (Reid et al., 1978; Reid, 1979;

Reid et al., 1979). Kiersite increased apparent Mg absorption and

retention (Reid et al., 1979), dry matter digestibility, live weight

gain, and feed intake of forages in sheep and pregnant beef cattle (Reid

et al 1978; Reid et al., 1979).

Concentrates contained variable amounts of Mg with highest amounts

found in the bran (.4 to .6% Mg, Reid, 1983). Magnesium bioavailability

in concentrates is high relative to forages (NRC,1978). Greater

bioavailability in concentrates may be due to smaller particle size

rendering Mg more available for digestion by rumen microorganisms.

Also, highly soluble, and readily available carbohydrates in concentrate

feeds enhanced Mg digestion and absorption (Wilson et al., 1969; House

and Mayland, 1976; Madsen et al., 1976). Wether lambs had sixfold

higher magnesium absorption and twofold higher urinary excretion when

sucrose was added to the diet, but no change occurred when comparable

amounts of starch were added (House and Mayland, 1976). Feeding 186,

373, or 932 kcal of glucose to sheep increased apparent absorption of Mg

linearly (Madsen et al., 1976), and drenching dairy cows with a starch

solution increased plasma Mg (Wilson et al. 1969). Field (1983)

suggested that concentrates lower ruminal pH and render Mg more soluble

in ruminal fluid, thus facilitating absorption.

Utilization of Magnesium by the Ruminant

Site and mechanism of absorption

Early studies characterizing absorption of Mg in ruminants concluded

that the small intestine was the primary site of absorption (Stewart and

Moodie, 1956; Field, 1961b; Storry, 1961a; Care and van't Klooster,

1965; Phillipson and Storry, 1965; Scott, 1965; Pfeffer et al., 1970).

In an early experiment with sheep prepared with isolated jejeunal and

ileal intestinal loops, Ca and, to a lesser extent, Mg were found to be

absorbed by the jejeunum and ileum (Scott, 1965). Experiments by others

demonstrated no absorption of Ca or Mg in the rumen but substantial

absorption of Ca from the upper and mid-jejeunum and concurrent, though

lesser absorption of Mg throughout the entire small intestine

(Phillipson and Storry, 1965). When sheep were cannulated in various

sections of the gut and blood samples taken from veins draining the

major digestive organs, Mg was absorbed from all sections of the

alimentary tract with greatest quantities absorbed from the small

intestine (Stewart and Moodie, 1956). Later experiments, where

absorption from the entire alimentary tract was investigated,

unequivocally demonstrated that the rumen was the primary site of

absorption (Grace and MacRae, 1972; Ben-Ghedalia et al., 1975; Strachan

and Rook, 1975; Tomas and Potter, 1976a,b; Field and Munro, 1977; Brown

et al., 1978; Martens et al. 1978; Fitt et al., 1979; MacGregor and

Armstrong, 1979; Greene et al., 1983; Martens, 1983). Absorption of Mg

in the ruminant was demonstrated in vitro using bovine ruminal

epithelium (Martens et al., 1978), in partial digestion studies using

cannulated sheep and steers (Tomas and Potter, 1976a,b), and in

ruminally fistulated dairy cattle in which Mg transport through the

ruminal wall was estimated (Martens, 1983). Although it appears that 70

to 100% of total Mg absorption can occur in the rumen (Tomas and Potter,

1976a,b; Greene et al., 1983) substantial amounts can be absorbed in the

small and large intestines (Field, 1961b; Pfeffer et al., 1970; Grace

and MacRae, 1972; Grace et al., 1974). In cattle, net absorption

actually was greater in the large intestine than small intestine due to

large amounts of endogenous Mg secretion which resulted in negative

values from the small intestine (Greene et al., 1983). There is limited

evidence that some absorption may occur in the omasum (Edrise and Smith,

1979). Allsop and Rook (1979) found in sheep infused with MgC12 that

true absorption of Mg was depressed when plasma concentrations were

greater than 3.0 mg/100 ml. Site of absorption may shift from rumen to

intestine depending on whether animals are fed continuously or once

daily (Grace and MacRae, 1972).

Magnesium absorption is saturated at high alimentary tract Mg

concentrations and has been associated with active transport (Martens et

al., 1976; Field and Munro, 1977; Martens et al., 1978; Martens, 1983).

Martens et al. (1978) using in vitro preparations of ovine ruminal

epithelium showed influx of Mg was temperature dependent, inhibited by

ouabain, and saturated at a ruminal buffer concentration of 5 mM Mg. A

similar type of phenomenon was reported by Brown et al. (1978) working

with sheep prepared with dorsal ruminal pouches in which Mg transport

was saturated at 4 mM Mg. In another experiment, Martens (1983) infused

a series of buffer solutions with varying Mg concentrations into washed,

empty rumens of dairy heifers, and demonstrated saturation of Mg

transport with 12.5 mM Mg. These data suggested that saturation

occurred at a lower concentration in sheep compared to cattle. Rate and

extent of Mg transport was related directly to the molar Na:K ratio in

ruminal fluid with greater absorption occurring as the ratio increased

(Smith and Horn, 1976; MacGregor and Armstrong, 1979; Martens and

Rayssinguier, 1980; Care et al., 1982). In sheep, maximum absorption

occurred at a molar ratio of 5:1 (Na:K) (Martens and Rayssinguier,


Field (1983) proposed a model for Mg absorption in the rumen in

which absorption is entirely from the liquid phase and dependent upon

extent of Mg solubilization in ruminal fluid. Adherence of Mg to the

surface of solid phase particulate matter was theorized to be

pH-dependent with significant amounts of Mg adhering above pH 6.0

(Field, 1983). Smith and Horn (1976) showed a dramatic decrease in

liquid phase concentration of Mg as ruminal pH rose above 6.0.

In contrast to the mature ruminating animal the small intestine is

the foremost site of Mg assimilation in calves (Smith, 1959a; Smith,

1962; Perry et al., 1967). In the dairy calf, 25 and 35% of Mg intake

was absorbed in the small and large intestines, respectively, and the

capacity of the large intestine to absorb Mg was diminished by 3 to 4 mo

of age (Smith, 1959b). By 10 to 28 mo, age had no effect on Mg

absorption presumably because the rumen is sufficiently developed.

Greater Mg absorption in sheep than in cattle may occur (Shockey et al.,


Relationships of sodium and potassium to magnesium absorption and

The adverse effect of K on Mg absorption has been studied

extensively (Kunkel et al., 1953; Fontenot et al., 1960; Kemp et al.,

1961; Suttle and Field, 1967,1969; House and Van Campen, 1971; Newton et

al., 1972; Tomas and Potter, 1976a; Green et al., 1983; Wylie et al.,

1985). This inhibition appears to be limited to the rumen (Care, 1965;

Tomas and Potter, 1976a). A linear decrease in Mg absorption was noted

when excessively high dietary concentrations of K were administered to

sheep and cattle (MacGregor and Armstrong, 1979; Greene et al., 1983).

These experimental dietary concentrations rarely would be encountered in

practical feeding systems. For example, in an experiment (Greene et

al., 1983) with steers, Mg absorption was diminished 39% when animals

were fed 4.8% dietary K or eight times the NRC requirement (NRC,1978).

Potassium also may have physiological effects at the kidney although

this is not as well documented. Intraruminal dosing of 1.5 g potassium

chloride (KCl)/kg body weight in nonpregnant beef cows increased net

tubular reabsorption of Mg (Deetz et al., 1981) and infusion of 33.6 g/d

K into rumens of lambs decreased urinary excretion of Mg from .28 to .11

g/d (Wylie et al., 1985).

Sodium may have a positive effect on Mg homeostasis by promoting Mg

absorption in the rumen. At the tissue level, use of ouabain or

dinitrophenol abolished Mg absorption which indicated Mg absorption is

dependent on adenosine triphosphate (ATP) synthesis and its presence in

the membrane (Martens and Rayssiguier,1980; Martens, 1985). Ruminal

concentration of Na may be critical to ATPase activity and subsequent Mg

absorption in the rumen. There are also indications from in vivo

experiments with sheep that Na enhanced Mg absorption in the gut (Mosely

and Jones, 1974; Huntington et al., 1977; Powley et al., 1977).

Metabolism in tissues

The mature animal does not have a readily available pool of Mg

because bone Mg is not readily accessible in periods of dietary Mg

deprivation (Parr, 1957). Tracer studies with both normal and

hypomagnesemic sheep indicated that rate of exchange of Mg between

plasma and bone is low, but relatively high for soft tissues (Field,
1960; Field, 1961b; McAleese et al., 1961). When 2Mg was injected

into mature sheep approximately 2.0% eqilibrated with bone undergoing

rapid metabolism (Field, 1960). In mature lactating cows, soft tissue

stores were depleted in lieu of bone magnesium in cases of chronic Mg

deficiency (Yoshida, 1981). In contrast, young animals rely upon bone

stores to supply metabolic requirements of tissues during Mg-deficiency

(Smith 1959b; Thomas, 1959) and in calves fed whole milk diets deficient

in Mg, bone Mg was reduced 30Z; serum Mg was lower also (Smith 1959c).

Metabolism: Excretion patterns

Of the total amount of Mg truly absorbed the portion excreted was

found primarily in urine and endogenous secretions in the

gastrointestinal tract (Smith, 1959b; Field, 1962; Pfeffer et al., 1970;

Grace et al., 1974; Ben-Ghedalia et al., 1975). Loss of endogenous

fecal Mg in the lactating dairy cow was approximately threefold higher

than urinary Mg excretion (Lomba et al., 1968). In dairy cattle

producing 10 to 20 kg of milk/d, 80% of total Mg ingested was excreted

in feces, and 12% and 3% were lost in urine and milk, respectively

(Hutton et al., 1965). In the ovine, an inverse relationship between

endogenous and fecal excretion and urinary excretion was reported

(Chicco et al., 1972). For example, animals which had lower fecal Mg

had greater urinary Mg so that net loss was a linear function of total

Mg intake (Chicco et al., 1972). Pattern of Mg excretion is entirely

different in the dairy calf compared to mature cows (Smith, 1959b).

Calves lost Mg primarily in urine, but with increasing age there was a

shift from urine to endogenous fecal losses (Smith, 1959b). Fecal Mg

was 32% in 3 wk old calves, and 86% in 16 wk old calves (Smith, 1959b).

This phenomenon also existed in sheep (Field, 1966).

Estimates of endogenous fecal Mg losses, which arise primarily from

gastric juices and saliva (Allsop and Rook, 1979), are limited for

lactating cows. Endogenous fecal Mg losses varied between .7 and 14 g/d

for nonlactating cows (13.4 mg/kg body weight) and 11.7 and 25.3 g/d for

lactating cows (33.6 mg/kg body weight, Lomba et al., 1968). Greater

losses by lactating cows were due to greater dietary intake and

absorption. Estimates of endogenous fecal Mg with sheep ranged from

approximately .7 to 5 mg/kg body weight (House and Van Campen, 1971;

Chicco et al., 1972; Allsop and Rook, 1979). In sheep infused

intravenously with MgC12, endogenous losses of Mg increased .0295 mg/d

for each increase of 1 mg/100 ml Mg in plasma (Allsop and Rook, 1979).

Diet did not influence endogenous secretions of Mg (Allsop and Rook,

1979). Generally, endogenous Mg losses were relatively fixed except at

excessive or very low plasma concentrations (McAleese et al., 1961;

Allsop and Rook, 1979) and urinary Mg reflects any increase in Mg


Urinary excretion is the most important mechanism for Mg

homeostasis. Magnesium in urine increased linearly with increased Mg

absorption in the lactating cow as the capacity for net tubular

reabsorption is exceeded. Magnesium status determined daily excretion

(Jesse et al. 1981). High producing dairy cows needed triple the amount

of Mg required by dry cows (Kemp, 1983). Renal responses to intravenous

infusion of Mg solutions occurred within minutes (Deetz et al., 1981)

and excretion of excess, highly available Mg from the diet or an

intraruminal dose occurred within hours (Deetz et al., 1981; Jesse et

al., 1981; DeGregorio et al., 1981). Comparing two MgO, the supplement

with greater laboratory reactivity had an excretion rate 61% higher than

the source of lesser reactivity (Noller et al., 1986). Also, urinary Mg

declined more slowly for the less reactive MgO after treatments were

terminated (Noller et al., 1986). Jesse et al. (1981) reported oral

dosing of MgO with smaller particle sizes were recovered quantitatively

the following day whereas larger particle sizes took 3 to 4 d to appear

in the urine. When lactating cows were fed supplements which provided

excessive dietary Mg for 70 to 90 d, urinary Mg declined indicating gut

absorption was saturated (Jesse et al., 1981).

Intraruminal infusion of citric acid increased plasma Mg clearance

and excretion in sheep (DeGregorio et al., 1981). In cattle,

intraruminal infusion of sodium citrate increased plasma Mg clearance,

glomerular filtration rate, and net tubular reabsorption of Mg (Deetz et

al., 1981). However, it is not clear whether the urinary responses were

due to Na or to the citrate ion since diets were not equalized for

either ion. Calcium was affected similarly (Deetz et al., 1981).

Although K increased the net tubular reabsorption of Mg, Mg salts

decreased K excretion in the ruminant (Wilson, 1964).

The Lactating Cow and Magnesium

Numerous experiments conducted over the past 10 years examined

effects of feeding MgO to dairy cattle. Interest in MgO was due to its

role in sustaining milk fat content under high concentrate feeding

systems for lactating dairy cattle (Emery and Thomas 1967; Huber et al.,

1969; Thomas and Emery, 1969a,b) although several researchers did not

find MgO to have this effect (Askew et al., 1971; Stout and Bush, 1972;

Gentry et al., 1978). Increases in milk fat percent with MgO frequently

occurred in cows fed 60 to 70% concentrate diets. Milk yields were

reduced in several of these trials; thus, it is unclear the extent to

which milk fat percent increased in response to reduction in total

yields. There was some evidence that MgO acted at the tissue level to

increase milk fat synthesis (Emery and Thomas, 1967). In a recent study

MgO had striking effects on milk production (Teh et al., 1985). In this

experiment early lactation cows were supplemented with .4 % or .8% MgO;

milk production was increased 9.8% in those fed .4% MgO with no response

in cows fed higher levels of MgO. Optimum response to MgO appeared to

be with .44 % dietary Mg. In another experiment milk yield responses

were not seen with MgO possibly due to excessively high dietary Mg (.56

to .7%, dry basis, Thomas et al., 1984).

Effects of MgO on milk yields and milk fat percent in these studies

were suggested to result from ruminal and(or) lower tract alkalizing

effects. In other experiments MgO increased ruminal pH, decreased

ruminal propionate and increased ruminal acetate concentrations (Thomas

and Emery, 1969a; Erdman et al., 1980; Jesse et al., 1981). It may

improve nitrogen and starch digestibility (Erdman et al., 1980; Erdman

et al., 1982). Thomas et al. (1984) found MgO increased fecal pH yet

there were no differences in starch (or dry matter) digestibility. In

addition, milk yield responses seen in the study of Teh et al. (1985),

associated with feeding .4% MgO, showed no changes in fecal pH, whereas

the higher Mg treatment (.8% MgO) demonstrated significant increases in

fecal pH but no milk yield responses. It is doubtful whether fecal pH

is an appropriate indicator of extent or rate of starch digestion. Most

likely, increases in fecal pH were due to undigested MgO reaching the

lower digestive tract (Teh et al., 1985).

Magnesium requirements for milk production synthesis in the high

producing cow have largely been neglected. The current NRC (1978) Mg

requirement is .2% of diet dry matter and is based on daily maintenance

needs of 6 to 7.5 g of Mg, and .12 g for each kg of milk produced,

assuming 33% bioavailability. The requirement of the ARC (1980) is

similar, though slightly lower, and is approximately 20 g/d Mg. Both

requirements were based largely on early experiments designed to

establish minimal daily requirements of mature animals for maintenance,

milk production and above all, prevention of grass tetany.

In Vitro Methods of Determining Magnesium Absorption by the Gut

Prior to advent of various in vitro techniques commonly employed

today, animal physiologists in the 1950s to 1960s studied Mg absorption

with surgically prepared animals (Stewart and Moodie, 1956; Phillipson

and Storry, 1965; Scott, 1965). These experiments were limited

basically to anesthesized sheep. The actual mechanism of Mg absorption

at the tissue level was studied by Martens et al. (1978).

Ussing Chamber Approach

The Ussing chamber technique for studying ion transport through

membranes by measurement of ion fluxes using radioactive tracers and the

short circuit current was first presented by Ussing and Zerahn (1951).

This approach was adapted to many tissue types, including amphibian

skin, urinary bladder, gills, anal papillae, insect rectum, intestinal

mucosa, salivary gland, mammary gland, kidney tubules, intestinal

mucosa, and ruminal epithelium (Keynes, 1969). The technique of

surgically preparing, mounting and sustaining ruminal tissue in the

Ussing chamber was reported by Ferreira et al. (1964,1972) and Stevens

(1964,1967). These experiments showed that ruminal tissue could be

maintained for several hours under in vitro conditions if the Ringer's

solution bathing the tissue was gassed continuously with 95% oxygen and

5% carbon dioxide and supplied with appropriate nutrients, such as VFA

and(or) glucose (Keynes, 1969). Estimates of viability of the tissue

generally were made by assessing the electrical potential. In vitro

preparations had slightly smaller but more stable electrical potentials

than those found in live animals (Keynes, 1969). In these experiments,

electrolyte, water, and VFA uptakes by the tissue were measured using

radiolabelled tracers.

Martens employed techniques worked out by Stevens (1964,1967) and

Ferreira et al. (1964,1972) to study Mg uptake by ruminal epithelial
tissue using Mg (Martens et al., 1976). In these experiments bovine

ruminal tissue obtained at slaughter was mounted in Ussing chambers with

Ringer's solution added to mucosal and serosal sides. Viability of the

preparation was monitored by measuring the electrical potential of the

tissue. Ruminal tissue was allowed to incubate for 2 h before the

bathing solutions were replaced and tracer added. Flux of Mg from

luminal to blood side was found to be 2 ug/(cm2. h). Transport of Mg

to the serosal side is in fact a nonphysiological process (Martens et

al., 1978). Magnesium was not taken up by the capillary blood system in

the ruminal villi; thus, there was a 25% accumulation of Mg in the

tissue. Both ouabain and dinitrophenol inhibited Mg absorption in this

system which indicated absorption was dependent upon active transport

(Martens et al., 1976; Martens et al., 1985).

Brush Border Membrane Vesicles

Brush border membrane vesicles have been prepared primarily from

intestinal and kidney brush border membranes (Stevens, 1984;

Ropfer,1978). Two techniques have been developed for preparation of

brush border vesicles. The first approach took advantage of the

rigidness of the brush border as compared with the basolateral membrane,

a property which allowed brush border membranes to be separated from

other cellular membranes by differing sedimentation rates (Murer and

Kinne, 1980). In the second approach the brush border was separated

from other cellular membranes by precipitation with Ca or Mg. The brush

border was found to be capable of binding divalent cations, a property

which allowed the membranes to be separated from other cellular

membranes (Murer and Kinne, 1980). Following precipitation with Ca or

Mg, vesicles formed by a series of ultracentrifugation steps. Vesicle

shrinkage was minimized by formation in the presence of inert solutes

such as mannitol or sucrose (Hopfer, 1978) and vesicle diameter usually

was calculated to be 100 to 150 nm (Stevens et al., 1984). Purity of

the membranes is assessed by enzymatic assays and vesicles were used in

transport experiments utilizing radiolabelled tracers and a rapid-

filtration technique.

Brush border membrane vesicles offered several advantages to the

study of membrane transport phenomena (Hopfer, 1978; Murer and Kinne,

1980). One advantage is the separation of membrane specific events from

cytosolic events. In many cases it was difficult to study transport if

the nutrient of interest was rapidly metabolized by the cell. The

experimenter is able to preset the conditions on both sides of the

membrane and study 1) direct versus indirect coupling ratios associated

with nutrient transport phenomena and 2) factors which affected

transport phenomena such as membrane potential difference and solute

concentrations (Murer and Kinne, 1980). Another advantage of this

system is it allowed the asymmetry of transport systems to be defined

(Murer and Kinne, 1980).

Disadvantages can include lengthy preparation time which possibly

caused inactivation of transport systems by alteration in permeability,

unstirred water layers, and changes in driving forces with time (e.g.,

abnormally slower or faster transport rates compared with those which

normally occurred in vivo) (Murer and Kinne, 1980). Other problems

which were cited were heterogeneity of vesicles, origin of membranes,

concentration of transport systems, leaky vesicles, and sidedness.

Sidedness did not appear to be a problem with brush border membrane

vesicles although it can be with basolateral vesicles (Hopfer, 1978).

Compared with the amount of data available on amino acid and glucose

uptake into brush border membrane vesicles there is little information

on ion movement per se (Hopfer, 1978; Fondacaro and Garvey, 1983).

Brush border membrane vesicles have not been used previously to study Mg

influx probably due to problems obtaining an adequate supply of 28Mg.

Membrane vesicles were used to study Na, Cl, iron (Fe), Ca and phosphate

ion influx. Brush border membrane vesicles have been used to study the

mechanism of Fe uptake into Fe-deficient and sufficient mice; in these

experiments Fe was suggested to be transported by a two-step process: 1)

Fe was translocated across the membrane by a carrier mediated process

and 2) binds to the inside of the membrane (Muir et al., 1984). This

system also was used to study the mechanism of Na and Cl translocation

in rabbit ileum (Fan et al., 1983). Brush border membrane vesicles were

used to demonstrate a Na-H antiport system in the rat small intestine

(Murer et al., 1976) and a ATP dependent H -K exchange mechanism in

gastric mucosa (Sachs et al., 1976; Chang et al., 1977).

Calcium binding and uptake in rat duodenal intestinal brush border

membrane vesicles were studied in detail (Miller and Bronner, 1981;

Miller et al., 1982). Time course experiments showed that Ca binding to

the brush border reached equilibrium within 5 min with .0092 mM Ca and

60 min with 3.6 mM Ca. Binding and uptakes were highest within the

range of 7 to 9 pH units and dependent on extravesicular Ca

concentration. With lower Ca concentrations (.0092 mM) maximum binding

occurred at pH 8.0 and with 3.6 mM Ca highest binding occurred at pH

9.0. Lower pH inhibited Ca binding, an effect which was overcome by

higher Ca concentrations. CalZ.uum was bound to brush border membrane

vesicles to both low and high affinity sites; the lower affinity sites

may have represented binding to phospholipid (Miller et al. 1982).

Calcium uptake was inhibited by extravesicular lanthanum, choline

chloride and ruthenium red (Miller and Bonner, 1981). Magnesium, barium

and Mn inhibited Ca binding to brush border membrane vesicles with

greatest effect seen for Mn (Miller et al., 1982). Replacing chloride

(Cl) with thiocyanate or nitrate ion increased Ca uptake by increasing

relative negativity of the vesicle interior (Miller et al., 1982).

Increasing extravesicular Na or K from 0 to 100 mM resulted in reduction

of Ca uptake from 5 to 2 nmol/mg membrane protein (Miller and Bronner,

1981). Calcium influx into vesicles from vitamin D3-deficient rats

was significantly less than into vesicles from vitamin D3 sufficient

animals (Miller and Bronner, 1981).

lonophores and the Ruminant

lonophores were defined as substances which enhance the movement or

incorporation of ions from the aqueous phase into the hydrophilic phase

(Houslay and Stanley, 1982). The word ionophore meant "ion bearer" and

was coined by Pressman (1976).

Mode of Action of lonophores in the Whole Animal

There are currently more than 70 known ionophores (Schelling, 1984).

Only two of these, monensin and lasalocid were approved for use in the

feed industry (Schelling, 1984; Goodrich et al., 1984). Effects of

lasalocid and monsensin were similar in the whole animal; other

ionophores/compounds which are known to have similar ruminal and whole

animal effects but are not currently approved are salinomycin, narasin,

and avoparcin (a glycopeptide antibiotic) (Schelling, 1984).

Ionophores were best known for their pronounced effects on rumen

fermentation. Richardson et al. (1976) demonstrated that monensin

increased ruminal propionate and decreased ruminal acetate

concentrations. Ionophores stimulated propionate and succinate

producing bacteria (Chen and Wolin, 1978), and inhibited lactate

producers in vitro (Dennis et al., 1981). Monensin had no effect on

lactate producers in vivo; yet it decreased acidosis caused by switching

animals to high concentrate diets (Burrin and Britton, 1986). This

effect of monensin was due to its action reducing total VFA rather than

changing lactate concentrations in the rumen (Burrin and Britton,

1986). In several experiments monensin decreased protozoa numbers, but

this was probably not one of its more important effects. Monensin

decreased rumen turnover rates of both solid and liquid material

(Schelling, 1984) and elevated ruminal pH which was related to decreased

total VFA (Burrin and Britton, 1986).

lonophores had systemic effects which were probably related directly

to ruminal action (Schelling,1984). Increased ruminal synthesis of

propionate caused increased propionate and decreased amino acid

utilization for gluconeogenis (Van Maanan et al.,1978; Beede et al.,

1980). In the lactating cow, the shift in the acetate:propionate ratio

caused decreased milk fat percent (Beukelen et al., 1984). In beef cows

decreased methane loss and increased propionate was partially

responsible for improved energy (ME) values (Goodrich et al., 1984).

Decreased proteolysis in the rumen was documented in animals fed

ionophores, which increased protein reaching the small intestine for

absorption (Poos et al., 1979). Summarizing six trials, the optimum

crude protein requirement of 11.6% for 246 kg steers was reduced to

11.2% for animals fed monensin (Goodrich et al., 1984). Lowest

feed:gain ratios in animals fed monensin were seen in steers fed

preformed protein as opposed to nonprotein nitrogen (NPN) (Goodrich et

al., 1984).

Other effects of ionophores included improved feed efficiency,

improved dry matter and nitrogen digestibility, improved feed to gain

ratio, and decreased feed intake (Goodrich et al., 1984), which were

probably related to ruminal action. Summarizing 228 trials with 11,274

head of cattle, monensin-fed cattle gained 1.6% faster, and ate 6.4%

less than control; feed requirements/100 kg gain were 7.5% less due to

monensin. These data indicated that 1) cattle which were inefficient in

gaining weight were more responsive to monensin than others, 2) better

responses were seen in cattle fed larger metabolizable energy intakes,

and 3) feed intake was improved in cattle which had (normally) lower

intakes but decreased in cattle with (normally) high intakes. Monensin

fed to growing steers increased apparent N digestibility and retention

18% and 46% respectively (Beede et al., 1986). Results from energy

metabolism trials indicated that monensin reduced maintenance

requirements, energy retention per unit of gain, and fasting heat

production; it improved net energy for maintenance to a greater extent

than net energy for gain (Goodrich et al., 1984).

Calcium ionophores had several overt physiological effects if

administered in pharmacological doses. Increased cardiac output,

increased contractivity of the heart, decreased peripheral resistance,

and decreased coronary resistance were demonstrated in the dog

(Pressman, 1983). These cardiovascular effects were mediated by Ca and

(or) catecholamines (Pressman, 1983).

Effect of lonophores on Mineral Metabolism and Absorption in the

In the past five years researchers evaluated possible effects of

ionophores on mineral absorption. Starnes et al. (1984) postulated

lasalocid or monensin fed at 33 mg/kg feed (dry basis) increased the

molar Na:K ratios in the rumen and may have been responsible for

increased absorption of Mg. Serum zinc (Zn) and copper (Cu) were

increased by monensin and lasalocid (Starnes et al., 1984). Doran et

al. (1986) fed .64 or .86% K, and 0 or 27 ppm monensin to feedlot cattle

and found no change in ruminal concentrations of Na or K; K had no

effect on in situ or N disappearance. Steers fitted with ruminal,

duodenal and ileal cannulae were fed 22 mg/kg feed monensin (dry basis)

or 30 mg/kg lasalocid (dry basis) to determine site of mineral digestion

(Gado et al., 1986). Ruminal P digestion was significantly lower for

steers fed monensin or lasolocid (17.6% for control versus 15.4 and

15.2%, respectively). Ruminal K disappearance was 25.1% for control,

20.9% for monensin, and 21.9% for lasalocid treatments (Gado et al.,

1986). Apparent digestibilites of Ca, P, Mg, and K in the small

intestine were lower in control versus ionophore diets by approximately

10% for P, Mg, K and 20% for Ca (Gado et al., 1986). Apparent total

digestibility of Ca was 8% higher for monensin and 12% higher for

lasalocid diets, respectively (Gado et al., 1986). Lasolocid increased

soluble ruminal concentrations of Mg, Zn, Fe and P and increased K in

red blood cells (Spears and Harvey, 1985). In lambs infused ruminally

with 0, 7.6, or 31.6 g/d K, and fed none or 20 mg/kg body weight

monensin, the ionophore increased apparent absorption and retention of

Mg 27.0 and 18.1%, respectively (Greene et al., 1985). Monensin did not

affect apparent absorption of Mg but increased Mg retention in lambs

52.% (Kirk et al., 1985). Apparent absorption of P increased 35.2% and

P retention increased 26.8% in monensin fed lambs (Kirk et al., 1985).

Apparent absorption of Zn increased 50.0% and Zn retention increased

45.0% by monensin (Kirk et al., 1985). Monensin also decreased liver Ca

45.4%, and bone Ca by 2.93%; urinary Ca decreased 60% in lambs (Kirk et

al., 1985).

The data on effects of ionophores on mineral absorption in ruminants

are inconclusive. In general, monensin and lasalocid appeared to have

positive effects on mineral absorption, but it is difficult to assign a

specific functional role of each ionophore in mineral absorption.

Several studies did indicate that Mg absorption or retention is

increased by monensin, and(or) lasalocid (Starnes et al., 1984; Greene

et al., 1985; Kirk et al., 1985; Gado et al., 1986; Oscar et al.,1986).

Mode of Action at the Cellular Level

lonophores range in molecular weight from 200 to 2000 daltons

(Houslay and Stanley, 1982) and are both naturally occurring and

synthetically derived; all have ion transporting capabilities but differ

in their method of transport and affinity for various ions (Pressman,

1976). In general, ionophores are ring structures or linear molecules

having a carbon backbone containing multifunctional groups (e.g.,

ether, alcohol, carboxyl, or amide). A single cation binds to

functional groups of the ionophore, and the carbon backbone of the

ionophore bends to accommodate the cation. Polar sections of the

complex are shielded from the solvent. In the case of the neutral

ionophores, the charge on the complex is the charge of the bound ion

(Pressman, 1976). For the carboxylic ionophores, the resulting complex

is a zwitterion (Pressman, 1976). Pressman (1976) has speculated that

ionophores may have originated from prosthetic groups of membrane bound

carriers; hypertrophy of these genes may have resulted in increased

numbers in bacteria giving some organisms a selective advantage.

Several classes of ionphores exist (Pressman, 1976). These include

the neutral ionophores, carboxylic, and channel forming ionophores.

Neutral ionophores include valinomycin, the enniatins, the

macrotetralide nactins and the synthetic polyethers (Pressman, 1976).

An example of a neutral ionophore is valinomycin, which is highly

specific for K. Some examples of carboxylic ionophores are monensin,

nigericin, X-537A (lasalocid) and A23187. Examples of channel forming

ionophores include gramicidin and alamethecin.

Ionophores are classified by both their structure and mode of ion

transport (Pressman, 1976). In the case of neutral ionophores such as

valinomycin, the ionophore diffuses through the membrane to the membrane

solvent interface where it encounters a cation for which it has a high

specificity. Water molecules surrounding the ion are removed and

replaced by oxygen molecules of the functional groups on the ionophore.

The complex next diffuses to the opposite side of the membrane where the

cation is released (Pressman, 1976). Rate of transport of the

cation-ionophore complex is determined by membrane potential and

concentration gradients on either side of the membrane (Pressman, 1976).

The mechanism of ion transport for carboxylic ionophores is

different from that of neutral ionphores (Pressman, 1976). The

carboxylic ionophores are open chained compounds with a terminal

carboxyl; they are capable of forming rings by hydrogen bonding head to

tail (Pressman, 1976). This type of ionophore is capable of being

protonated at the carboxyl group; in this form it diffuses through the

membrane to the solvent membrane interface (Pressman, 1976). The proton

is released and the charged ionophore is believed to be trapped proximal

to the polar solvent environment (Pressman, 1976). The ionophore binds

a cation, making the complex now a zwitterion, and diffuses to the

opposite side of the membrane where the opposite sequence of

microequilibria reactions takes place (Pressman, 1976). Transport is

dependent on chemical concentration, not membrane potential since there

is an electroneutral exchange of a cation for a proton (Pressman, 1976).

The channel forming ionophores have ion transporting capabilities

entirely different from the neutral and carboxylic ionophores (Houslay

and Stanley, 1982). There are two classes of channel forming

ionophores. One class is the polyenes, of which amphotericin and

nystatin are members. They are composed of conjugated lactone rings and

are similar to phospholipids in their size and structure (Houslay and

Stanley, 1982). One side of the molecule is hydrophobic and can

interact with cholesterol in the membrane; the other side is very

polar, having many hydroxyl groups (Houslay and Stanley, 1982). Pores

are formed in the membrane by multiple amphotericin molecules aligned in

the membrane with the nonpolar sides of lactone rings interacting with

cholesterol; this forms a polar channel having an interface of several

hydroxyl groups (Houslay and Stanley, 1982). The polarity and size of

the channel permit nonspecific passage of a variety of hydrated ions and

anions (Houslay and Stanley, 1982). Gramicidin is a member of the

second class of ionophores, a class consisting of peptide ionophores

(Houslay and Stanley, 1982). Gramicidin has a left handed helix of

hydrophobic amino acids (Houslay and Stanley, 1982). To form a pore,

two molecules insert in the membrane, head to tail with hydrophobic

amino acid residues aligned with membrane phospholipids; carboxyl groups

of the amino acids are orientated toward the middle of the channel and a

polar corridor is formed (Houslay and Stanley, 1982). These ionophores

form smaller pores and hence are more specific than the first class of

channel forming ionophores.

Ionophores vary in their specificity (Pressman, 1976; Houslay and

Stanley, 1982). Some such as valinomycin are highly specific.

Valinomycin binds K with 103 to 104 times greater affinity than Na

(Houslay and Stanley, 1982). Sodium has a larger desolvation energy

and is less stable since it is smaller (diameter .19 nm) than K

(diameter .27 nm). The Ca transporter A23187 is highly specific for

divalent cations due to the presence of two N atoms in the interior ring

structure (Pressman, 1976). Two molecules of A23187 bind to every Ca

molecule (Pressman, 1976). Other ionophores such as A-537A (lasalocid)

are less specific. Lasalocid complexes with virtually every cation;

this is thought to occur because ions sit on the molecule, rather than

bind within it (Pressman, 1976). The pore forming ionophores can be

cation specific or relatively nonspecific depending on the size of their

pore (Houslay and Stanley, 1982).

There is some limited information on effects of ionophores on

eukaryotic and prokaryotic cells. In mouse fibroblasts, salinomycin,

monensin, and lasalocid stimulated Na uptake and K extrusion (Austic and

Smith, 1980). Stimulaton of ATPase was 3 to 4-fold and was indicated by
active uptake of 86Rb in the cells; effects of ionophores on the

membrane appear to be reversible (Austic and Smith, 1980). Monensin

stimulated alpha-aminoisobutyric acid accumulation by Swiss 3T3 cells.

Since uptake was inhibited by ouabain it was concluded that monensin

stimulates the Na-K pump in the plasma membrane (Austic and Smith,

1980). In avian red blood cells monensin stimulated uptake of the

glucose analog 3-0-methyl-D-glucose due to increased Na uptake by the

cells which in turn stimulated mitochondrial uptake of Na and extrusion

of Ca (Bihler et al., 1985). Uptake of the analog by avian red blood

cells was hypothesized to occur via a depletion of intracellular Ca

(Bihler et al., 1985). Monensin has been shown to cause accumulation of

epidermal growth factor in lysosomes presumably because it depletes the

hydrogen ion pool, and high acidity (within lysosomes) is necessary for

degradation of the ligand (King, 1984). In prokaryotes, effects of

rumen microbial physiology appear because they are toxic to rumen

organisms, in particular to gram positive bacteria (Bergen and Bates,

1984). Toxicity results from their ability to decrease the proton

motive force, deplete cellular ATP, and dissipate the proton gradient

(Jolliffe et al., 1981).

lonophores are used in membrane vesicle systems to establish a

specific electrical gradient across the vesicle membrane by means of

salt gradients (Kaunitz and Wright, 1984; Fan et al., 1983). Available

data on ion transport as affected by ionophores in brush border

membranes is limited. Miller and Bronner (1981), using rat duodenum,

demonstrated rapid accumulation of Ca in the presence of the putative Ca

ionophore A23187 (15 uM); at 30 min intravesicular Ca in the presence

and absence of the ionophore is virtually the same. Calcium release

effluxx) by the vesicles in the presence of 15 uM A23187 and .1 mM EDTA

was accomplished in 2 min (Miller and Bronner, 1981; Miller et al.,




Potassium (K) has an inhibitory effect on absorption of magnesium

(Mg) in the rumen (Tomas and Potter, 1976a; Greene et al., 1983).

Evidence for this inhibition includes a high incidence of grass tetany

in cattle grazing pastures with K content greater than 3.0% (dry basis)

(Kemp, 1983), and experiments where apparent absorption of Mg decreased

when animals were infused intraruminally with K or fed high K diets

(Tomas and Potter, 1976; Greene et al., 1983). How K elicits this

effect has not been determined; it may decrease absorption of Mg by

altering the electrical potential of ruminal epithelial tissue (Tomas

and Potter, 1976a). Although the adverse effects of K on ruminal

absorption of Mg have been studied, limited data exist on effects of

sodium (Na) on ruminal absorption of Mg (Smith and Horn, 1976; Powley et

al., 1977; MacGregor and Armstrong, 1979; Martens and Rayssiguier,

1980). Smith and Horn (1976) suggested that a high molar ratio of Na:K

in ruminal fluid promoted ruminal absorption of Mg. Sheep grazing grass

with high K content were in negative Mg balance until allowed to graze

pastures high in both Na and K (Powley et al., 1977). In vitro

experiments with ruminal epithelial tissue demonstrated that Mg was

absorbed from the ruminal liquid phase and that flux was dependent upon

the molar ratio of Na:K in the perfusing buffer (Martens and

Rayssiguier, 1980). The effect of the molar ratio of Na:K was studied

in sheep by perfusing washed empty rumens with a series of buffers

varying in Na and K concentrations. Results demonstrated raising the

molar ratio from 1:1 to 5:1 increased absorption of Mg from .5 to 1.5

mM/h (Martens and Rayssiguier, 1980).

In our experiment the primary objective was to determine effects of

high K and Na inputs on apparent disappearance (absorption) of Mg from

the rumen of the lactating cow. We studied ruminal disappearance of Mg

and other macrominerals by measuring changes in liquid outflows of Mg

from the rumen as ;a indication of apparent disappearance (absorption)

because Mg, and possibly other ions, are absorbed from the liquid phase

(Field, 1983). We assumed that if outflows of minerals from the rumen

were altered then apparent absorption was changed also, providing daily

liquid phase outflows were not altered due to treatments.

Materials and Methods

Animals and Design

The experiment was an randomized incomplete block design with eight

cows, eight treatments and five periods. Eight ruminally

fistulated-Holstein cows in late lactation were kept in tie-stalls for

the entire experiment (35 d) and were milked twice daily at 0700 and

1900 h. Diets were offered ad libitum at 0800 and 1800 h. Water was

available continuously and consumption by individual cows was measured

by in-line flow meters. Periods were 4 d in length. A fifth period was

used to estimate single period carryover (residual) effects thus each

cow remained on the treatment assigned in period four during period



Composition and structure of experimental treatments are in Tables

3-1 and 3-2. Treatments were arranged as a 2 X 2 X 2 factorial with .20

or .40% Mg, .87 or 1.65% K and .22 or .60% Na. Treatments were selected

for the following reasons. The lower mineral concentrations were near

NRC (1978) recommendations for each ion of interest. The upper dietary

concentrations were those studied previously at the University of

Florida where significant lactational responses were noted (Schneider et

al., 1984; Schneider et al., 1986). Treatments were constructed from

feed mineral and intraruminal administrations of either Mg acetate

(reagent grade, Fisher Scientific Co., Orlando, Fl), potassium chloride

(KC1) (feed grade, International Minerals Co., Mundelein, II), or Na

tripolyphosphate (reagent grade, Fisher Scientific Co., Orlando, Fl).

Intraruminal administration of minerals was to insure that cows received

appropriate amounts of highly soluble and available minerals for

potential absorption. Mineral salts were dosed intraruminally twice

daily at 0800 and 1700 h. Cows on the lower Mg treatment received 29 g

Mg daily from the basal diet plus 38 g/d from Mg acetate. This was

equivalent to either .20 or .40% total Mg (dry basis). Potassium and Na

treatments were constructed similarly. Cows on the low K treatment

received 138 g/d of K from the basal diet (.87%, dry basis); those on

the high K treatment received an additional 116 g/d of K via

intraruminal administration of KCI (1.65%, dry basis). For the Na

treatments the basal diet provided 31 g/d of Na (.22%, dry basis); the

high Na treatment consisted of basal dietary Na plus 67 g/d of Na from

Na tripolyphosphate administered intraruminally (.60%, dry basis).

0uC0000 rm c
0 w0 m lA01 m 'r-g I -
. 4 I I I .
0 u4l -4 -4 -4
eL UrN
Oc mmu\ti


m Lr

) 00ac
0* co C
oa LA LA -
en *




0 O 0
0 CM 0

0 0
0~ 0

C) n n
M in

0 0 C-1 C# V)

C)* 0

00 r-n

0. 0 1 en r I

m D1000 M

1ee00 e

* o r
I A 0O





0t0 emmoeeme
*0r-' m0 0 o-m
. .-. .

n- -

4 '

0 co U) a
-) -4 0 4-1tu" 0

C: cc m x 0
p 4) W ) $ a
edea e a i x
0W mU
0 r 0C .W0 3
04 W -4 e U W

Sa) (U o a 1
.0 CO = P O P0 C
0 11 1 *. *- *4l-l 0 4 E 0 0
u~w&s i>Q &.

-4 -) ..


** It

* *
--<-4 -4


* I

r'-I II
'I I I




G -4
* *~

-4 Co

I I *-4



U 0.
U 0j W

a) -4o
w wH C

u 0 0

S41 1-i
0 0 0

4 04 4)

S0 0 U
ac oin w U

U )

-4 0

C 0
u a
















0 .



S )


4 w
*r 1



0 0

o o

0 0

























0 a

u w

E-t 4
C 0

** *
I '0'00 0 '0> 0'0<

* *

4 0 0 t40




IIIIl Il lt


I I I I I Il t
I l l I N I U I 0

* *

a *


I I M I I I 4 I

I I I I I I j I I
I I 1 P

I I I I I.

I l l lio'0 Po 0

- N PM I P M

I 0h 0< CPI 0% Ch 0, 0%








I 0% 0 N cN

1 -tM^4CYf<^




t l




H 3


S 0




0 u

c c

1 .1




4U N

Hl 0

C o 4.)

H 0


The basal diet was approximately 57% ground corn, 30% cottonseed

hulls, and 5% soybean meal plus minerals and vitamins (dry basis) (Table

3-1). Diets were constructed to promote relatively low ruminal pH

because liquid phase Mg adsorbs onto solid phase ruminal ingesta above

pH 6.0 (Smith and Horn, 1976). Experimental diets were formulated to

balance intraruminal administration of minerals so that all anions,

cations and nonprotein nitrogen were equivalent except for Mg, K, and

Na. Potassium chloride and potassium sulfate were used as sources of K

because the chloride and sulfate components could be balanced with

ammonium chloride and ammonium sulfate in the basal diet. Urea was

added to balance nonprotein nitrogen from ammonium chloride and ammonium

sulfate. Dietary phosphorus (P) was equalized using dicalcium

phosphate. Limestone addition equalized calcium (Ca) among treatments.

Acetic acid (1% v/v) was administered intraruminally to cows not

receiving Mg acetate to equalize input of acetate.

Sampling and Analyses

Ruminal contents were collected on the fourth day of each period.

At 0800 h, 2.7 g of chromium (Cr) as a 1 liter Cr ethylenediaminete-

traacetate (EDTA) solution (Binnerts et al., 1968) was administered

intraruminally to all cows and throughly mixed in ruminal contents.

Immediately after Cr-EDTA administration an initial sample of ruminal

contents was removed from each cow (time=0) and ruminal samples were

taken every 3 h for the next 24 h. There were eight samples of whole

ruminal contents collected per cow each period. Immediately after

sampling, the liquid phase was separated from the solid ingesta by

filtration through two layers of cheesecloth yielding approximately 400


0: II 0 0 -

w ,-4 co U 4M

CC--4 -- *
ol* +

4 U-4

0 0 -

) %- + .+ .
- 0 0) i* 1. *

0 C) m

c 1; C4J
- *i.. e m o o -

- z H m OD r 01

0 C M

Q (0

U Cd -
r%4 Z

--- 0 .0 o

Scc b -- 0 V 0
w w

04 Ed -00 0
.0 0
4 M 0

W -4 -, o o

>1 0* 0
u4: ao

J4-i = 3+ +
E-i Q3 L c s E;t! + 4;

ml of liquid and 500 g of ruminal solids. Ruminal fluid pH was measured

immediately after separation. .Both fractions were frozen immediately

(-10 C). Feed intakes and milk yields were measured daily. Individual

milk samples were taken daily and analyzed for fat and protein (Dairy

Herd Improvement Laboratory, Raleigh, NC).

Feed samples were ground in a Wiley mill (2 mm), composite within

treatment and a subsample used for analyses (New York State Forage

Testing Laboratory, Ithaca, NY). Ruminal solids were composite within

cow-period, dried in a forced-air oven (100 C) for 48 h, and ground in a

Wiley Mill (2 mm). Samples were wet-ashed digested with nitric acid and

analyzed for Mg, Na, K, and Ca concentrations by atomic absorption

spectrophotometry (Model 5000, Perkin-Elmer, Inc., Norwalk, CT).

Trihourly ruminal liquid samples from each cow were analyzed for Cr

concentrations by atomic absorption spectrophotometry. Samples were

pooled within cow-period and mixed in a high capacity commercial

blender. As the pooled sample was blending, a 10 ml subsample was

collected for macromineral analyses. Pooled samples (within cow-period)

for Mg, K, Na and Ca analyses were deproteinated with 10%

trichloroacetic acid, 1% lanthanum solution and analyzed by atomic

absorption spectrophotometry. Phosphorus concentrations in ruminal

fluid, rumen solid, and plasma samples were analyzed by an ascorbic

acid-ammonium molybdate procedure (Technicon Industrial Method 94-70WB)

with a Technicon II Autoanalyzer (Technicon, Terrytown, NY). Plasma

samples were deproteinated and processed similarly to ruminal fluid

samples for mineral analyses.


Liquid phase dilution rates were obtained by calculating the slope

of the line of the plot of the natural logarithm of ruminal fluid Cr

concentrations versus time. Ruminal volumes were estimated by dividing

the antilogarithm of the dependent variable (Cr concentration) intercept

of this plot by the quantity of Cr administered. Daily outflows of

ruminal liquid phase were calculated as the product of ruminal volume

times dilution rate times 24 h. Outflows of Mg, K, Na, Ca and P in the

liquid phase were estimated by multiplying cow-period means of liquid

outflows by concentrations of minerals in ruminal fluid.

Statistical Analyses

Data were analyzed by method of least squares analysis of variance

using the general linear model procedures of the Statistical Analysis

System (SAS, 1982). The original mathematical model for analysis of

variance contained treatment, period, cow, and carryover effects.

Carryover effects were not detected for any response variable;

therefore, the final model contained treatment, period, and cow as

sources of variation.

Results and Discussion

Results of dietary main effects on dry matter and water consumption,

ruminal fluid volume, dilution rate and ruminal liquid outflow are in

Table 3-3. Higher Na increased water intake approximately 10 liter/d

(14% increase) (P<.10). When water intake data were analyzed using dry

matter intake as a continuous independent variable (covariate), there

was no effect of Na, indicating that increased water intakes actually

were resulting from increased dry matter intakes with high Na

treatments. Other researchers found that added dietary Na increased

water intake when dry matter intake was held constant (Rogers and Davis,

1982a) or ad libitum (Rogers and Davis, 1982b). In the study presented

here higher K increased water intake 13 liter/d (18% increase; P<.10).

Significant effects of K on water intake remained when dry matter intake

was used as a continuous independent variable (covariate) in the

mathematical model (P<.05). Ruminal infusion of K in sheep

significantly increased water intake (Tomas and Potter, 1976a). Ruminal

fluid volumes of our cows were similar to values reported by others

(Rogers and Davis, 1982a; Erdman et al., 1982). There were no

significant main effects of treatments on ruminal fluid volumes.

However, there were two significant (P<.05) two-factor interactions.

Magnesium and K interacted to increase fluid volumes whereas Na and K

interacted to decrease fluid volumes. Similar ruminal fluid volumes

resulted with .87% K, and either .20 or .40% Mg, but increased 28% with

.40% Mg versus .20% Mg with 1.65% K (Mg X K interaction) (P<.05). With

.87% K, ruminal fluid volumes were 18% greater with .60% Na versus .22%

Na, but 10% lower with 1.65% K with .60% Na versus .22% Na (Na X K

interaction) (P<.05). Rogers and Davis (1982a) found ruminal fluid

volumes were increased 14% by intraruminal infusion of 288 g sodium

bicarbonate (NaHCO 3) and 200 or 600 g sodium chloride (NaCI) in


Least squares means for dilution rates ranged from 8.6 to 8.9%/h

similar to other reports (Rogers and Davis, 1982a; Rogers and Davis,

1982b; Erdman et al., 1982). In several studies, high mineral intakes

by lactating cows (Schneider et al., 1987) and steers (Rogers and Davis,

1982) increased ruminal dilution rates. In our experiment, ruminal

dilution rates were quite consistent and not affected by treatments.

There were no significant main effects of treatments on daily ruminal

liquid outflow. However, there was a significant K X Mg interaction on

daily ruminal liquid outflow which resulted from the identical

interaction on ruminal volumes because dilution rates were not changed.

Liquid phase outflows decreased 21% as K increased from .87 to 1.65%

with .20% Mg, but with .40% Mg, outflows increased 14% as K increased

(Mg X K interaction; P<.10). Daily ruminal fluid outflows were similar

to values obtained in experiments with lactating cows fed higher dietary

macrominerals (Erdman et al., 1982). In steers, total ruminal liquid

outflows were increased by administering NaCl and NaHCO3 (Rogers and

Davis, 1982a).

Although the length of periods in this experiment were too short to

evaluate treatment effects adequately on dry matter intake and

production, we noted increases due to treatment. Least squares

treatment means for milk yields ranged from 13.3 to 15.5 kg/d. Higher

Mg increased milk yields 17% (P<.05). There were no detectable effects

of Mg, K, or Na on milk fat percent. Potassium decreased milk protein

percent slightly (3%). Schneider et al. (1986) previously reported

increased dietary K decreased milk protein percent. Higher Na increased

dry matter intakes 2.1 kg/d (15% increase; P<.05), whereas additional Mg

increased dry matter intakes 2.4 kg/d (18% increase; P<.05). Higher K

did not affect dry matter intakes, in contrast with Schneider et al.

(1986) but in agreement with Schneider et al. (1984). The Na effect was

consistent with Schneider et al. (1986) who found .55% Na (dry basis)

from NaCI or NaHCO3 improved dry matter intake in longer-term studies

with lactating cows. Short-term intake responses to magnesium-deficient

diets were reported (Chicco et al., 1973a; Ammerman et al., 1971).

Voluntary intakes were diminished greatly in the ovine within 24 h if a

Mg-deficient diet were offered and immediately improved if a

Mg-sufficient diet were given (Chicco et al., 1973a; Ammerman et al.,

1971). In these experiments decreased feed intakes apparently resulted

from decreased cellulolytic activity in the rumen (Chicco et al., 1973a;

Ammerman et al., 1971) although low cerebral spinal fluid Mg

concentrations may modulate the intake response (Seoane et al., 1973).

Studies with sheep and beef cattle demonstrated that fertilization of

forages with Kiersite (Mg sulfate) increased forage Mg and dry matter

intake in both sheep and cattle (Reid, 1983).

Dietary main effects on ruminal pH, molar ratios of Na:K, ruminal

liquid and solid phase ion concentrations and daily ruminal liquid phase

ion outflows are in Table 3-4. Values for ruminal pH found in this

study (5.6 to 6.0) were expected to promote absorption of magnesium by

maintaining liquid phase magnesium ultrafiltrable (Storry, 1961b).

Sodium increased ruminal pH .3 units (P<.01). This suggested that

effects of NaHCO3 bicarbonate to raise ruminal pH (Rogers and Davis,

1982b; Teh et al., 1985) may be due partially to alkalizing effects of

the Na ion. Least squares means for molar ratios of ruminal Na:K ranged

from 1.7 to 3.5 for individual cows. These estimates were similar to a

report where molar Na:K ratios in rumens of steers ranged from 1.5 to

6.0 (Smith and Horn, 1978). In sheep, Mg absorption increased linearly

as the ratio increased from .4 to 3.0; maximum absorption of Mg occurred

0 CM 0
M >0 0 0 000
* 0 M.
'0 4 4% C
CMI -^

. I

O -





P4 H

44 F~O
94 PA


OD (m

4 0




g0 0
N^ C

04 @.NN
* ** *
SfN 0% '0
m c

00 0iUo
m IMm -
oN CM -
0 N


. 0


eg m



I g

o -

"3 -

.4 o
I j

*0 N

0 0

o 4


o o

U -

0 N

0 0
0* '. 0



0 .0

0 P 0

4 '

0 aA 0

N 0


U a

.M 00
a -


01 ^

* *a
3* .4
JI m


0f'4 000
00P. 0-4f'
0%L 0% '0

.0 P.

0 N 0%

P.. -
0 sA


fN -

Go U


4 a



4a N

0 0

0 N 0

N P.
M0 0

0 S0
4 N

P '

a0 N

0 .

UL 0

1. -

.4 U




v "
a -
0.I. I


C 04.
M 0

U U .-3.
- !04

0 W% 0
mn CN a,
o -


4' a

C V-

U 04
H 0

2 09 .2
S- U

: I -f
0 o-4
0. cn

0 n 0 0 r- 0
CM o o r-


0 *M 0. 0 a;
4 a

000 0 0

P(4 n C4 In

fn a *A r
(' U' P

m o

0O 0

0M 4
04 (A
0 u


0o V

0 0
'0 '
o _

V a

Cu H 0 + ac 0 K

with a molar ratio in the range of 3 to 5 (Martens and Rayssiguier,

1980). In our study higher K decreased molar Na:K ratios from 3.2 to

1.8 (P<.001). It has been observed that as dietary K increased, it

replaced Na in saliva and lowered this ratio in the rumen (Martens and

Rayssiguier, 1980).

In our study, we expect that maximum absorption of Mg occurred

because Mg concentrations in ruminal fluid ranged from 190 to 360 ppm or

8 to 15 mM. In experiments with heifers ruminal absorption of Mg was

maximum at 12.5 mM (Martens, 1983). Least squares means for Mg

concentrations in ruminal fluid increased 90% with the higher Mg

treatment (P<.001). There was a significant Na X Mg interaction on

ruminal liquid phase concentrations of Mg (P<.10). With .22% Na and

.20% Mg ruminal fluid concentrations were 183 ppm and increased to 197

ppm with .60% Na; with .40% Mg, ruminal Mg concentrations were 390 ppm

and decreased to 330 ppm as Mg increased from .22 to .60%. Solid phase

concentrations of Mg represent average Mg content in ruminal solid

samples taken every 3 h over a 24-h period. Solid phase Mg

concentrations were 150 ppm with .20% Mg and increased to 270 ppm with

.40% Mg (P<.001). Compared to Na and K dietary treatments which did not

influence solid phase minerals, Mg administration greatly increased

solid phase Mg (80% increase). This may have occurred due to binding of

Mg to cottonseed hulls in the rumen. High Mg treatment increased Mg

liquid phase outflows from the rumen 70% (P<.001). There were no

effects of K administration on outflows of liquid phase Mg. Dietary

concentrations of K needed to inhibit absorption of Mg are higher

generally than 2.0% (dry basis) (Greene et al., 1983). There was a

significant Na X Mg interaction on Mg outflow from the rumen (P<.10).

Magnesium liquid phase outflows were not affected by higher Na with .20%

Mg, but decreased 21% with .40% Mg and .60% Na. The interaction was due

entirely to decreased ruminal liquid phase concentrations of Mg, with no

concurrent increase in solid phase Mg content. Data suggested an

enhanced absorption of Mg due to higher Na, because decreased ruminal

outflows of Mg were not due to decreased liquid outflows or transference

of liquid phase Mg onto the solid phase.

Ruminal concentrations of K agreed with those of Schneider et al.

(1987) and increased 37% with higher K (Table 3-4) (P<.001). Magnesium

administration increased ruminal concentrations of K 11% (P<.10). There

were two significant two-way interactions on ruminal liquid phase

concentrations of K. With .20% Mg and .22% sodium, ruminal K

concentrations were 1652 ppm and increased to 2091 ppm with .40% Mg and

.22% Na; K concentrations decreased from 1807 ppm with .20% Mg and .60%

Na, to 1762 ppm with .40% Mg and .60% sodium (Na X Mg interaction,

P<.05). As Mg administration increased from .20 to .40%, ruminal

concentrations of K increased from 1361 to 1729 ppm with .87% K; with

1.65% K ruminal liquid phase concentrations of K increased from 2099 to

2133 ppm as Mg increased (K X Mg interaction, P<.10). Higher K

increased daily liquid phase outflows of K 32% (P<.05). Outflow of K

from the rumen decreased 20% with .40% Mg as Na increased to .60%, but

increased 22% with .20% Mg as Na increased to .60% (Na X Mg interaction,

P<.10). This suggested Na and Mg interacted to increase K absorption in

the rumen.

Ruminal concentrations of Na were similar to those of Schneider et

al. (1987) and decreased 17% with higher K (P<.05). Solid phase Na

concentrations were reduced 12% by higher Mg (P<.10), 16% by higher K

(P<.01), and increased 15% by higher Na (P<.05). Potassium

administration decreased liquid phase outflows of Na 33% (P<.001), due

to decreased Na concentrations in ruminal fluid, liquid phase outflow

was not changed due to K administration (Table 3-3). Others have noted

that under conditions of high K intake there was increased absorption of

Na from the rumen (Rook and Balch, 1958). Also, K may have replaced Na

in saliva (Kemp and Geurink, 1978). Ruminal fluid Na concentrations

decreased 50% when sheep were infused intraruminally with KCI (Tomas and

Potter, 1976a).

All Ca and P data were analyzed statistically with dry matter intake

in the model as a continuous independent variable (covariate) because Na

and Mg both significantly (P<.05) increased dry matter intake (Table

3-3) and hence Ca and P intake. Ruminal calcium solid and liquid phase

concentrations and flow were analyzed with carryover effects in the

model since this was a significant source of variation (P<.05). Higher

Na decreased ruminal concentrations of Ca 28% (P<.05). This caused a

29% reduction in Ca liquid outflows suggesting Na promotes absorption of

Ca in the rumen. Higher Na did not affect ruminal liquid outflow (Table

3-3, P>.10). Ruminal liquid phase concentrations of P decreased 19%

with higher K (P<.01). Potassium administration decreased ruminal P

outflows 21% which indicated K increased P absorption in the rumen

(P<.05). Higher K did not alter ruminal liquid phase outflow (Table

3-3, P>.10). This is consistent with findings of Greene et al. (1983)

who found 2.4 and 4.5% dietary K (dry basis) increased preintestinal

absorption of P approximately 40% in steers. In our study higher Na

increased P concentrations in ruminal fluid 9% (P<.01).

In summary, we evaluated administration of higher quantities of Mg,

K, and Na on ruminal kinetics and mineral disappearances in lactating

cows. At the concentrations used in this study, alterations in ruminal

volumes, dilution rates and outflows due to mineral treatments did not

occur. Dietary mineral concentrations used in this study frequently

occur in diet formulations. For example, NaHCO widely used as a

buffer, increases dietary Na (Schneider et al., 1936; Rogers and Davis,

1982a,b), whereas K contents of forages of 2 to 4% (dry basis) are not

unusual (Ward, 1966). Additionally, use of MgO as a dietary buffer can

raise dietary Mg from .2% to .6% (dry basis) (Thomas et al., 1984).

There appeared to be no adverse effects of 1.65% K on Mg absorption in

our study. Our results did not unequivocally support the proposal that

the molar ratio of Na:K in ruminal fluid dictates extent of magnesium

absorption. We found that higher K significantly lowered the ratio. It

would follow that if the ratio was important, K would have had a

significant negative impact on Mg disappearance from the rumen, which

did not occur. Our study supported the idea that Na, and not the ratio

per se, was important in governing extent of Mg (and Ca) absorption in

the rumen.


Water intakes increased 14% and 18% with higher dietary Na and K,

respectively. No effects of treatments on liquid dilution rates were

detected and there were no main effects of treatments on ruminal liquid

outflows. However, there was a significant K X Mg interaction; with

.20% Mg liquid phase outflows decreased 21% as K increased from .87 to

1.65%, but with .40% Mg outflows increased 14% as K increased. Higher

Na increased ruminal pH. Higher K decreased molar ratio of Na:K in

ruminal fluid from 3.2 to 1.8. Higher Mg increased daily Mg liquid

outflows 70%. Ruminal concentrations of Mg were 183 ppm with .22% Na

and .20% Mg, and increased to 197 ppm with .40% Mg; ruminal Mg

concentrations were 390 ppm and decreased to 330 as Na increased from

.22 to .60% (Na X Mg interaction). There was a significant Na X Mg

interaction on ruminal outflow of Mg; Mg outflows were not affected by

Na concentrations with .20% Mg, but decreased 21% with .40% Mg and .60%

Na. Results suggested no effects of K on ruminal disappearance of Mg,

and enhanced ruminal disappearance (absorption) of Mg due to Na.



In previous experiments we noted production responses to diets

containing approximately 1.50% potassium (K) (dry basis) and .40 to .60%

sodium (Na) (Schneider et al., 1984; Schneider et al., 1986). These

concentrations were considerably higher than current NRC (1978)

recommendations. We studied apparent ruminal absorption of

macrominerals in Holstein cows administered mineral treatments equal to

.20 or .40% magnesium (Mg), .87 or 1.65% K, and .22 or .60% Na (dry

basis)(Chapter 3). In this chapter we present effects of these

treatments on renal processing of Mg, K, Na, calcium (Ca), and

phosphorus (P).

Limited information exists on effects of Na and K upon urinary

excretion of Mg by lactating dairy cows. There was a linear

relationship between Mg intake and urinary excretion of Mg in excess of

the renal threshold of approximately 2.0 mg/100 ml (Kemp, 1983). A

renal threshold for Mg apparently exists because Mg net tubular

reabsorption decreases as the filtered load increases (Wong et al.,

1983). Magnitude of daily urinary loss was dependent upon Mg intake and

milk production (Kemp, 1983). There is no compelling evidence for renal

secretion of Mg (Raynaud, 1962; Murdaugh and Robinson, 1960). Higher

dietary K decreased urinary excretion of Mg (House and Van Campen, 1971;

Newton et al., 1972; Deetz et al., 1981). Administration of K to

ruminants decreased renal excretion of Mg by increasing its net tubular

reabsorption (Deetz et al., 1981).

Sodium increased renal clearance of Ca in humans and sheep (Kleeman

et al., 1964; King et al., 1964; Tomas et al., 1973). Sodium also may

increase renal Mg loss in the ruminant (Burt and Thomas, 1961; Tomas et

al., 1973; DeGregorio et al., 1981). In sheep, drinking .8 or 1.3%

saline significantly increased urinary Ca and Mg excretion (Tomas et al.

1973). Effects of Na citrate on urinary Mg excretion were studied in

ruminants because of its speculated role in the etiology of grass tetany

(Burt and Thomas, 1961; DeGregorio et al., 1981). In cattle,

intraruminal administration of Na citrate (1.5 g/kg body weight) caused

higher Mg clearance (Deetz et al., 1981). In sheep, citric acid

increased Mg clearance and reduced net tubular reabsorption of Mg

(DeGregorio et al.,1981). This was believed to occur by completing of

the citrate anion with Mg and preventing net tubular reabsorption.

However, in these experiments treatments did not contain equal Na, thus

diuresis of Mg cannot be attributed directly to either Na or citrate.

Treatments containing higher concentrations of macrominerals in our

study (.40% Mg, 1.65% K, .60% Na, dry basis) are common in practical

diet formulations. For example, ruminal buffers, such as Na bicarbonate

(NaHCO3) and Mg oxide (MgO) elevate dietary Na and Mg contents (Erdman

et al., 1980, 1982; Thomas et al., 1984). Moreover, it is not unusual

for forages to contain 2 to 4% K (dry basis) (Ward,1966). The objective

of our experiment was to evaluate effects of intraruminal administration

of Mg, K, and Na on macromineral excretion by the kidney.

Materials and Methods

Experimental Design and Treatments

The experiment was an incomplete randomized block with eight

animals, eight treatments and five, 4-d periods. The treatment

arrangement was a 2 X 2 X 2 factorial with .20 or .40% Mg, .87 or 1.65%

K and .22 or .60% Na. Treatments were constructed from basal diet

minerals and intraruminal administration of either Mg acetate (reagent

grade, Fisher Scientific Co., Orlando, Fl), potassium chloride (KC1)

(feed grade, International Minerals Co., Mundelein, II), or Na

tripolyphosphate (reagent grade, Fisher Scientific Co., Orlando, Fl).

Intraruminal administration of mineral salts was twice daily at 0800 and

1700 h.

Animals and Diets

Eight late lactation Holstein cows were used. Cows were kept in

tie-stalls for the entire experimental period (35 d) and were milked

twice daily at 0700 and 1900 h. Diets offered ad libitum were fed at

0800 and 1800 h. Water was available continuously and individual cow

consumption was measured by in-line flow meters. The basal diet

consisted of approximately 57% ground corn, 30% cottonseed hulls, and 5%

soybean meal plus vitamins and minerals (dry basis) (Table 3-1). Diets

were constructed so that all dietary cations, anions and nonprotein

nitrogen were equalized with the exception of Mg, K, and Na (Table 3-2).

Sampling and Analyses

Whole blood was collected via the caudal vein from each cow on d 4

of each period at 1400 h into three, 10 ml ammonium-heparinized tubes.

Blood samples were centrifuged at 3000 X g for 20 min and the plasma was

stored at -20 C. Plasma was analyzed for Mg, K, Na, and Ca by atomic

absorption spectrophotometry (Model 5000, Perkin-Elmer Norwalk, CT).

Phosphorus was determined by the ascorbic acid-ammonium molybdate

procedure (Technicon Industrial Method 94-70WB) using a Technicon

Autoanalyzer II (Technicon, Terrytown, NY). Cows were fitted with

indwelling urethral catheters (Foly, 22 Fr with 75 ml ovoid, C.R. Bard,

Inc., Murry Hill, NY) 1 h prior to initiation of urine collection.

Total urine excreted over a 24-h period was collected on the last d of

each period into 20 liter carboys. Every 12 h on the last d of each

period, urine was weighed and a 500 ml sample was taken and stored at

-20 C. Also, urine samples (10 ml) were taken directly from the

catheter (after voiding catheter residual volume) twice daily for

immediate measurement of urine pH. Urine samples were analyzed for

minerals similarly to plasma samples. Creatinine concentrations in

plasma and urine were measured by a picric acid colorimetric method

(Henry et al., 1974).


Daily excretions of minerals were calculated as products of 24-h

urine output multiplied by urinary mineral concentrations. Fractional

excretions were calculated as [(urine mineral concentration/plasma

mineral concentration) '/'(urine creatinine concentration/plasma

creatinine concentration)] X 100 (Lane and Merritt, 1983; Thomas et al.,

1984). Renal mineral clearances (ml/min) were calculated as [(

concentration of urinary mineral X urine output/unit time)'/' plasma

mineral concentration)] (Deetz et al., 1981). Glomerular filtration

rates were estimated similarly using creatinine concentrations (Deetz et

al., 1981). Filtered loads of minerals were calculated by multiplying

glomerular filtration rates by plasma mineral concentrations (Deetz et

al., 1981). Net tubular reabsorption rates of minerals were estimated

as differences between excretion and filtration rates (Deetz et al.,


Statistical Analyses

Data were analyzed by method of least squares analysis of variance

using the general linear model procedures of the Statistical Analysis

System (SAS, 1982). The initial analysis of variance contained

treatment, period, cow and carryover effects as sources of variation.

The final mathematical model had treatment, period, and cow effects

because single period carryover effects were not detected. Higher Mg

and Na both significantly increased dry matter intake (Table 3-3), thus

it was appropriate to analyze Ca and P data with dry matter intake in

the mathematical model as a continuous independent variable (covariate)

to adjust intakes of Ca and P among cows.

Results and Discussion

Least squares means for urine pH, glomerular filtration rate, and

daily urine output are in Table 4-1. Although bovine urine is normally

alkaline having an average pH of 8.23 (Osbaldiston and Moore, 1971), it

was lower in our experiment (least squares treatment means ranged from

5.3 to 8.1). Administration of each cation had an alkalogenic effect on

urine pH. Higher Mg increased urine pH .6 units (P<.01) whereas, higher

K and Na increased urine pH 1.6 (P<.001) and .8 units, respectively

(P<.001). In experiments where higher dietary concentrations of sodium

chloride (NaCl) were fed, urine pH was reduced (Bailey, 1978; Schneider


0 M 4O -

W *

SC4 m

I 0O 0 I

H3 o N **
u.9g .;

a 00 n

U t

>I 1 N N *
0 4 -'

m *

S a
z w -i co QO

CZ 4 m C

<0 *** Wj

A4 0 C
0 ) 0

Sp oo o
0 E-< 1 co m >


4 0 ca 0
w4 w
0 l* *
3 o ~ c

SO *0) $ V
0) 2) ) C
00 0 0 V P.

M- E 0 91

et al., 1987), possibly due to high chloride (Cl) or exchange of Na for

hydrogen ion in the distal tubule (Pickering, 1965). Least squares

treatment means for glomerular filtration rates ranged from 534 to 1186

ml/min. These values were higher than those found for sheep which were

approximately 75 ml/min (DeGregorio et al., 1981) but, are similar to

rates for cattle of 800 to 1000 ml/min (Bailey, 1978; Pickering, 1965).

There was a trend towards higher K increasing glomerular filtration rate

(P<.11). In sheep, glomerular filtration rate was increased by infusion

of KC1 at a rate of 1.2 mmol/min (Scott, 1969a) or by feeding 889 to

1124 meq K/day (Scott, 1969b). In contrast, another study found

intraruminal infusion of 1.6 g of K/kg body weight did not increase

glomerular filtration rates in sheep (DeGregorio et al., 1981). In

cattle, increasing dietary NaCI from .5 to 7% caused glomerular

filtration rates to increase from 719 to 893 ml/min (Bailey, 1978). In

the current experiment, higher Na increased daily urine output 47%

(P<.001). There were two significant two-factor interactions on urine

output. Urine output increased from 24.4 kg/day to 30.1 kg/d with .87%

K as Na increased from .22 to .60%; with 1.65% K, urine output increased

from 17.2 to 30.8 kg/d as Na increased from .22 to .60% (Na X K

interaction, P<.10). Urine output increased from 23.5 to 27.2 kg/d as

Na increased from .22 to .60% with .20% Mg; with .40% Mg urine output

increased from 18.2 to 33.7 kg/day (Na X Mg interaction, P<.05).

Sodium, Na X K, and Na X Mg effects on urine output remained

(P<.01,P<.05, P<.10, respectively) when water consumption was included

in the mathematical model as a continuous independent variable

(covariate), indicating a diuretic effect of the minerals not associated

with altered water intake.

In Table 4-2 are least squares means for main effects of Mg, K and

Na on plasma concentrations, urine concentrations, daily excretion,

fractional excretion, filtered load, clearance and net tubular

reabsorption of Mg, K, and Na. Least squares treatment means for plasma

and urine creatinine concentrations ranged from 1.23 to 1.73 mg/100 ml

and 51 to 87 mg/100 ml, respectively. Higher Mg increased plasma Mg

concentrations from 2.3 to 2.7 mg/100 ml (Table 4-2). Higher Na

decreased urinary Mg concentration 27% (P<.05) possibly due to the

diuretic effect of Na which caused urine output to increase. Least

squares treatment means for daily Mg excretions (3.6 to 7.9 g/d) agreed

with other experiments in which daily Mg excretions varied from 1 to 8

g/d in lactating cows (Storry and Rook, 1963). Daily Mg excretion

increased 43% by higher Mg (P<.10). Daily excretion of Mg increased

111% as Mg increased from .20 to .40% with .87% K, but decreased 7% as

Mg administration increased to .40% with 1.65% K (Mg X K interaction,

P<.05). Least squares treatment means for fractional excretion of Mg

ranged between 16.9 and 33.8% similar to values of Thomas et al. (1984)

who found fractional excretion of Mg in lactating cows ranged from 14 to

31%. Fractional excretion of Mg decreased 27% with higher K (P<.10).

Clearances of Mg were approximately 30 to 50% higher than those reported

by Deetz et al. (1981) for beef cattle. Magnesium clearances increased

from 122 to 194 ml/min as administered Mg increased from .20 to .40%

with .87% K, whereas clearances decreased from 172 to 136 ml/min with

1.65% K (Mg X K interaction, P<.10). Deetz et al. (1981) did not detect

any differences in Mg clearances due to intraruminal infusion of KCI in

cattle. In our experiment, there was a trend towards increased net

-000 0 ui

4- 4 m


o oo o c
N N *O O

0 0 0 t 0 .

N 4 IC v; r-

0000 0 N0

e4 o O
c r^tC u



CCM or, CI



co4 o ee o
N M no co
t40 0 in


NInN 0
UCM .4o







o z



U l-


0PI 0 0 0 0 0

C0 0 CM 00

4'4N -

0 0 0 0 00
s el0 -.4-c
-40 N C C

o" -
0 00 4 CC 000

000 N 00 0
0 aN
coin 11 i .4'o. W

ob o
000 000


u 0 -0 F- o0

I Is z
-- -

C ....0 I.

0 a.C 0 U -k-. V I

in 4 >u
*r rf*>Jwu
*O ^ ki a l^ ^
Oa.obk Z
0} )cl

o o
0 0 -

.- c a c
S H 0 M H 0

C 1*. 0 0 LC 0 0

a s***4 L* cn we**4 j o -
'ou o- o u S0t

0 I- 5. ) 0 I.'

eU-U .
0 o0w 0 A C A. u 3
*- oc -u c

50.b~b~b 3 .UEZ O. 0

* -

- 0 *
* V
V V 4

tubular reabsorption of Mg by the kidney due to higher K (46%) (P<.11).

Deetz et al. (1981) reported intraruminal KCl (1.5 g/kg body weight)

increased net tubular reabsorption of Mg 46% in young cattle and 69% in

older cows. Other workers reported K diminished urinary Mg excretion in

ruminants (House and Van Campen, 1971; Newton et al., 1972).

There were no dectectable treatment effects on plasma K

concentrations in agreement with Schneider et al. (1984; 1987).

Potassium is almost totally reabsorbed in the proximal tubule, with

secretion occurring in the distal nephrons (Pickering, 1965). Urinary

concentrations of K were increased 129% by K (P<.001) and decreased 24%

by Na (P<.05). Administration of added K increased daily urinary K

excretion 84% (P<.01). Fractional excretion of K was increased 68% by

higher K administration (P<.05). Higher K increased K clearances 104%

(P<.05). Increases in K excretion due to K administration may have been

due partially to increased glomerular filtration rates. Higher K

increased net tubular reabsorption of K (P<.10). There was a

significant Na X Mg interaction on net tubular reabsorption of K

(P<.10). Net tubular reabsorption of K increased from 2.0 to 2.6 g/day

as Mg increased from .20 to .40% with .22% Na; with .60% Na, net tubular

reabsorption decreased from 6.2 to -.9 g/d as Mg increased from .20 to

.40%. Approximately 50% of filtered K was excreted in agreement with

results of Pickering (1965).

There were no main dectectable effects of treatments on plasma Na,

in agreement with Schneider et al. (1984; 1986; 1987). Sodium

administration increased urinary Na concentrations 109% (P<.001), daily

excretions 196% (P<.001), fractional excretion 135% (P<.001), and

clearance of Na 184% (P<.001). As suggested by Bailey (1978), who noted

similar effects of high Na intake in cattle, this indicated that Na

excretion increased at a faster rate than total urine excretion. Least

squares treatment means for fractional excretion of Na ranged from .1 to

2.7%, in agreement with Thomas et al. (1984) in which fractional

excretion of Na in lactating cows ranged between .61 and 2.3%.

Magnesium administration increased fractional excretion of Na 40%

(P<.05). Fractional excretion of Na increased 58% as Na increased from

.22 to .60% with .87% K, whereas with 1.65% K fractional excretion

increased 320% (Na X K interaction, P<.10). Reabsorption of Na was

quite high, yielding almost complete retention of Na by the kidney in

agreement with results of Pickering (1965).

In Table 4-3 are Ca and P excretion measurements. Calcium data were

highly variable. There were no significant treatment effects on

measurements of renal processing of Ca. There was a trend towards K

having diminishing effects on Ca excretion similar to its effects on Mg

excretion. Higher K tended to decrease daily excretions of Ca (41%)

(P<.33), fractional excretion of Ca (66%) (P<.35), and increased net

tubular reabsorbed Ca (42%) (P<.13). These trends agreed with reports

in which K depressed urinary Ca excretion (House and Van Campen, 1971;

Newton et al., 1972). Magnesium increased plasma P from 6.5 to 7.7

mg/100 ml (P<.05). Values for urine concentrations, daily excretions

and fractional excretions of P tended to be high in this experiment.

Higher values found in our study may be due to the nature of the diet,

as increased P excretion in sheep was associated with feeding high

concentrate diets which produce an acidic urine (Scott, 1972).

Ln t I-

0 ***. or^i t

- -"


r- C 000

LA O ; 8 1





u- C00 00 00


D-I CM 00

0 o\Cfw C4

O 0











< 0
! 3

o j

CM4 r- O-i o' i r

0 %D a %0 Ln M

0 o -4 I e

-i e


0 0 0 e
o *-4- o<->-

e a

U Q 0 C C.

cc *4 0 0
0 00- C 0C

0 4- X 13 .r. c

60 c U -. 0 "
0 '-4 W I 4 C8 1I -4
cc u :D = a) 0 0

e) O ) 1
aI C .- U .4-1 0


D- -4 C .

Ir I- CM 0 4



.1 4C

0) 0

-0r 4 -4 4
w u- c M 0

0 00 0

S0 60 U O. 0

0 4-1 .- $4 ., ,- 4-
o n)-B

0 0 .O 4 .. 0 --- ( 4

Additionally, the use of a highly available P source in our study (Na

tripolyphosphate) may have caused P excretions to be higher. Least

squares treatment means for urinary excretion of P, which ranged from 12

to 23 g/d, were higher than estimates of Osbaldiston and Moore (1971)

who reported that daily P excretion in cows in various stages of

lactation ranged from 2.5 to 6.5 g/d. In sheep and calves daily P

excretion ranged from 0 to 8 g/d, and as a percent of total P intake,

from 3 to 60% (Scott, 1972). In our study, P excretion in urine was

approximately 17% of daily P intake. Daily P excretion increased 79%

with .20% Mg as Na increased from .22 to .60%, whereas it decreased 12%

with .40% Mg as Na increased from .22 to .60% (Mg X Na interaction,

P<.10). There were no significant effects on fractional excretion of P.

What emerges from this study are the striking effects of K on Mg

excretion in the lactating cow. We found no inhibitory effect of K on

ruminal absorption of Mg (Chapter 3) as has been detected by others

(Tomas and Potter, 1976a). This suggests that at 1.65% dietary K no

inhibition of Mg absorption would occur, and this dietary concentration

would promote Mg retention by the animal provided it is not excreted by

an alternative route. Magnesium retention by the kidney under

conditions of excessive dietary K intake, found in some grazing

situations, may actually allow the animal to conserve Mg.

The ruminant kidney has the capacity to excrete copious amounts of K

(about 50% of the filtered load), and minimal amounts of Na (1 to 10% of

the filtered load) (Anderson and Pickering, 1962; Pickering, 1965).

When fed large amounts of NaCI, the ruminant also has the ability to

eliminate Na very efficiently (Bailey, 1978). In effect, the ruminant

kidney has adapted to dietary conditions of low Na and high K found in

many forages.


Higher Mg, K, and Na intakes compared to NRC (1978) increased urine

pH; higher Na increased urine output. Glomerular filtration rate was

increased slightly by higher K. Higher Mg increased daily Mg excretion

43% compared with .20% Mg. Higher K decreased fractional excretion of

Mg 27%. Daily Mg excretion increased 111% as Mg increased from .20% to

.40% with .87% K but with 1.65% K, daily Mg excretion was reduced 7%

with .20 compared with .40% Mg (Mg X K interaction). Clearance of Mg

increased 60% as Mg increased from .20 to .40% with .87% K, whereas, it

decreased 21% with 1.65% K as Mg increased (Mg X K interaction). Higher

K increased net tubular reabsorption of Mg 46%. Higher Na increased

daily excretion, fractional excretion and clearance of Na. Higher K

increased daily excretion, fractional excretion, clearance and net

tubular reabsorption of K. There were trends toward similar effects of

K on Ca excretion as were found for Mg.



Few studies have evaluated the dietary magnesium (Mg) allowances for

lactating dairy cows. Rook et al. (1958) verified that cows producing

approximately 10 kg of milk daily were in positive Mg balance if

supplied with 20 to 30 g Mg. Kemp et al. (1961) summarized data from 11

balance trials with Friesian cows (average daily milk yield 15 kg) and

concluded that they required 2.5 g/d of available Mg for maintenance and

.12 g/kg milk produced. In another early study, Blaxter and McGill

(1956) calculated that lactating cows producing 5 to 23 kg of milk daily

required 7 to 15 g Mg, assuming mean availability of dietary Mg of 33%.

Based on these experiments and several other studies, the ARC (1978)

published Mg requirements and an allowance factor for lactating cows.

For Friesian cows producing 10 to 40 kg of milk daily requirements

ranged from 10.4 to 23.1 g Mg whereas daily allowances ranged between

17.9 and 40.0 g Mg. In the U.S., the NRC (1978) suggested a requirement

of .20% Mg (dry basis) and recommended .25% Mg for lactating cows under

feeding or environmental conditions conducive to hypomagnesemia.

These suggested requirements and (or) allowances raise several key

points. Firstly, they are based largely on experiments with low

producing cows (10 to 20 kg/d), and may not reflect Mg needs of higher

producing animals. The cow has a considerable demand for dietary Mg for

milk production (Storry and Rook, 1962). Cows producing 25 to 45 kg of

milk daily secrete 3 to 6 g Mg in milk. Moreover, the Mg content of

milk does not decline when cows are switched to dietary regimens having

the potential to cause tetany (Rook et al., 1964). Secondly, these

requirements and (or) allowances were established primarily for

prevention of grass tetany as opposed to optimum dietary concentrations

for maximum milk production. Thirdly, they assumed average availability

of dietary Mg. In the case of ARC (1980) requirements, Mg availability

was assumed to be 29.4%. It is well established that Mg availability is

highly variable in typical feedstuffs for lactating cows (Kemp et al.,

1961; Rook and Campling, 1962). Kemp and Geurink (1978) pointed out

that a cow which produced 25 kg milk needed 55, 28, or 18 g Mg daily

depending on whether availabilities of dietary Mg were 10, 20, or 30%.

It is noteworthy that supplying the NRC (1978) requirement (.20% Mg)

oftentimes can be achieved simply from the Mg in feedstuffs in the diet;

however, if the availability of Mg in these feeds is low, it is doubtful

whether cows will receive their actual daily requirement of available

Mg. Based on these observations and a recent study in which feeding

higher dietary Mg (.44% Mg, supplied from magnesium oxide) increased

milk yield 9.8% (Teh et al., 1985), it was proposed that both current

ARC (1980) and NRC (1978) requirements and (or) recommendations are

inadequate to supply the daily Mg needs of higher producing cows.

In our experiment we used feed grade magnesium phosphate (MgP,

Boliden Kemi AB, Helsingborg, Sweden, 24% Mg, 13% P, .2 to 1.5 mm

granules), a product which has been available in western Europe for

several years. There are several ways this Mg source differs from

magnesium oxide (MgO): 1) MgP, which is made by reacting ground calcined

magnesite with phosphoric acid, is generally a more chemically

consistent product than most MgO sources and 2) studies in Scotland

demonstrated MgP in general has higher biological availability.

Hemingway (1985) concluded availabilities of four MgP sources were 34.9

+ 5.2% (mean + standard deviation) in contrast to 24.7 + 11.3% for 20

commercially available samples of MgO.

Our experiment was designed also to evaluate possible sodium (Na)

and potassium (K) effects on production responses as well as optimum

dietary Mg for lactating cows. Schneider et al. (1986) found that

during summer in Florida actual milk yields (not adjusted for fat

content or dry matter intake) increased 6.3% with .55% total dietary Na

supplied from sodium bicarbonate (NaHCO3) compared with .18% dietary

Na, and increased 4.2% as dietary K increased from 1.3 to 1.8%. Milk

yields (adjusted for dry matter intake but not fat content) increased

5.3% when dietary Na was increased from .18% to .55% with sodium

chloride (NaC1) (Schneider et al., 1984, 1986).

Objectives of this experiment were to 1) determine the optimum

dietary Mg allowance for lactating cows in mid-lactation, 2) assess

effects of dietary Mg, Na, and K and their possible interactions on

production responses and 3) assess treatment effects on renal fractional

excretions and blood plasma mineral concentrations.

Materials and Methods

In January, 48 mid-lactation Holstein cows were assigned in a

partially balanced randomized incomplete block design to three of 16

dietary treatments arranged in a 4 X 2 X 2 factorial: 1) .26, .38, .48

or .60% Mg, 2) .24 or .62% Na and 3) 1.14 or 1.59% K (dry basis) (Table

5-1). Dietary sources of Mg, K, and Na were endogenous minerals in

feedstuffs plus supplemental MgP, potassium bicarbonate (KHC03,

Schuylkill Chemical Co., Philadelphia, PA), or NaCl. Each cow received

three different dietary treatments in each of three 35-d periods. There

were nine cow-period observations per dietary treatment in the

experiment (144 total cow-period observations). Treatments were

arranged such that no treatment followed another treatment in a

subsequent period more than once during the entire experiment.

Composition of the basal diet is in Table 5-2. On a dry matter

basis, diets were 50:50, corn silage:concentrate. To formulate other

dietary treatments additional MgP, KHCO and(or) NaCI replaced ground

corn. All dietary treatments were equal in phosphorus (P) and calcium

(Ca) content; this was accomplished by varying concentrations of

dicalcium phosphate and limestone (Table 5-1). We assumed the

availabilities of P in dicalcium phosphate and MgP were equal based on

results of Hemingway (1985) who reported availabilities of 73% and 71%,

respectively. Cows were fed the basal diet for 1 mo prior to the start

of the experiment. Individual mineral premixes were pre-weighed for

each 909 kg concentrate mix made as needed throughout the experiment.

Total mixed rations were made by combining corn silage and concentrate

immediately prior to feeding (once per day) in a mobile mixing and

feeding unit with electronic scales (American Calan, Inc., Northwood,

NH). Ad libitum feed intakes and refusals were monitored daily for

individual cows using electronic gate feeders (American Calan, Inc.,

Northwood, NH). Cows were housed in groups of 12 in a free-stall barn

and allowed access to dirt lots adjacent to the barn at all times.

Drinking water was available continuously in the barn.



-4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4
* 0C C 0 CC












-4 C 00
* -4
-I -I I-



4 4 4 4 4 4 4
4 4 4 4 4 4 4 4~




---4-4- 4 -4-4-4~

44 4 0 0 (

C4 C

-4--4-4 -4 -44-4-4_-q

0 Q



0 -4
m 0



04 Z

, o

*4 E

0c- ,

O -4



4 0

.4 e

60 .-4
Ha >
c 0

60 to


0 IC

0 O

4 C 4

O -

4e >


0 CO -c'
c ^i cc

60 W

0a 0

u 4. cV) 0
C 4
W c
* o *


in.4 4) 0
O -4 C
+ -

4) W0 0

.-o 74 g1
O -N

0 0 -i

cc -4 o
U) ,W

a 0

U U >

B 0 0 80
W -

*E 4 a -*

~ a)a*

9 O0

m m wwo000 -44 -4 mC) m 0 0w 4It4-4

0% 0 0 00 0 0 0 0 00
o .00 .. 00

00 00 ON -4





U w
*-4 0



0 U
0 -


w 0










W -4

*i A


TABLE 5-2.


Ingredient Percent of dry matter SD

Corn silage 50.0
Ground corn 22.6
Soybean meal 10.0
Distillers dried grains with solubles 15.0
Trace mineralized sodium chloride .35
Urea .50
Limestone .80
Dicalcium phosphate .75

Laboratory Analysis

Crude protein, % 18.7 .2
Acid detergent fiber, % 19.2 .3
NE Mcal/kg 1.66 .001
Calcium, % .77 .04
Phosphorus, % .62 .01
Magnesium, % .26 .003
Potassium, % 1.14 .07
Sodium, % .24 .01
Iron, ppm 332 15
Zinc, ppm 56 6
Copper, ppm 8 1
Manganese, ppm 43 2

aMeans + standard deviations (SD) for each analytical component of the
basal diet were calculated from analyses of composite of samples taken
over the entire experiment for each 16 dietary treatments, with the ex-
ception of magnesium, potassium, and sodium which were calculated from
analyses of dietary treatments 1, 2, 9, and 10 for magnesium, 1, 2, 3, 4,
5, 6, 7, and 8 for sodium and 1, 3, 5, 7, 9, 11, 13, and 15 for potas-

Net energy for lactation calculated from formulation.

Concentrate samples for laboratory analyses were collected once

weekly, dried (60 C), composite within period and ground (Wiley mill, 2

mm screen). Corn silage samples were taken two to three times a week

for dry matter analysis (60 C) in order to adjust and maintain

proportions of corn silage and concentrate in the total mixed rations on

an as-fed basis. Dried corn silage samples were composite within

period and ground for later chemical analyses. All feed samples were

analyzed for nutrient content (New York State Forage Testing Laboratory,

Ithaca, NY). Milk samples for fat, protein and mineral analyses were

collected at morning and evening milkings the last 3 d of each period

and sent to the DRIA Testing Laboratory (Raleigh, NC) for fat and

protein analyses. Milk minerals were wet-ash digested and analyzed by

atomic absorption spectrophotometry (Model 5000, Perkin-Elmer, Inc.

Norwalk, CT). Average daily milk yields and feed intakes for individual

cows during the last 14 d of each period were calculated and used in

statistical analyses.

Urine and caudal or jugular vein blood samples were taken 4 to 6 h

after feeding from all cows on the last day of each period. Urine

samples were collected by vulval stimulation into plastic bags

containing 10 ml sulfuric acid and frozen (-10 C) for later creatinine

(Henry et al., 1974) and mineral analyses by atomic absorption

spectrophotometry. Blood samples were collected into heparinized

(lithium salt) tubes and immediately centrifuged (3,000 X g); plasma was

frozen (-10 C) for later analyses of minerals.

Data were analyzed by method of least squares analysis of variance

using general linear model procedures of SAS (SAS, 1982). Tables 5-3,

5-5, and 5-7 detail mathematical models. Treatment effects were tested

by orthogonal contrasts: linear, quadratic and cubic effects for Mg,

linear effects for Na and K, and all two and three-way interactions were


Results and Discussion

Intake, Milk Yield and Composition Responses

Least squares analyses of variance for feed intake and milk

production responses are in Table 5-3 and main effects least squares

(marginal) means are in Table 5-4. There was a linear (P<.05) effect of

Mg on dry matter intake with equal intakes with .26, .38, and .48%

dietary Mg and a 4.9% decline with .60% dietary Mg. The feed intake

response tended to be curvilinear (P<.10). Because all dietary

treatments were equivalent in P (.62 + .01%, mean + standard deviation),

it was assumed the decline in feed intake above .48% Mg resulted from

added dietary Mg as opposed to P. Thomas et al. (1984) found dry matter

intakes were 12 to 16% lower with diets containing .50% MgO or .55% to

.70% magnesium hydroxide (Mg(OH)2) compared with intake of cows fed no

supplemental Mg source. Others found MgO depressed dry matter intakes

(Thomas and Emery, 1969; Stout et al., 1972). Yet results of many

experiments indicated no effect of MgO on dry matter consumption (Erdman

et al., 1980; Jesse et al., 1981; Schaefer et al., 1982; Teh et al.,

1985; Erdman et al., 1982). In these experiments it is difficult to

determine whether effects of MgO on dry matter intakes were due to

possible gastrointestinal tract alkalizing effects of MgO or Mg contents

of the diets, which were .60% of dry matter or greater in most cases.

Several workers reported Mg fertilization of pastures, or dietary





E-4 Z
P 0


0 w



U >-

|X (


.4 y u

0.u4-4 ISeuuez

o -


0 00



I 0 0

I 0


N 0

I -


-oor-on oooooo-on~o

w000 0000000 4 00% 0

4 0 0 m I W% N m m f4 0

v v
. ..0044..' 0 N N 0 0
000,000 00000000


vV V
++ e

aP N r-0 fo- m Um m 0 m 0%
A N 0000 '- -----



_ ^ _._ _1_1_ 4-h


O 44
4 >

0 0
4 o o.
4 0

) o 0

4-4 44
r1 02 UC

0 00

a C

0 > s a)

0 rX4 ( )
)) rz r C

4-4 4) 0 0
4O 0 0 4J

1 -H 0 "l
0 0 0 r C

S0H 0 0

< a n a a m e
1 1 00 30

H) 44 0 t 00
0 00 N 0 u
Q) oi 44 0
S to H -4-4 44
U a u 14 0U 2 W )
0 4 0 H H b H H

4 4 CO 4- H14 *C O
0 3 4-4 % 3
C H 4-1 o U o

cc 0 L) H U
34 4-4 02 1 00 B0

S14 M 0 S 02 V P
Ca 43 a H4 l
0) 3 0 U. 0 U

0o a 0 4 o O
H O l OO 14 3 _
.( H C0 4 ^ cl 64 0 H n *



o o0. --

o o-- m
L4 U6 m oa N go

N4 Ne
0 0 0 a. 0 eor- a,

on iM en m

co4 c a N

o a ? a a-
PD m 0 OD en OD

00 c000
000 r- 00 W4

Na in N<*



E- Cl


g M~


-H 0

oa u
0 V V

0* v) 2*
Z 0) 0
4U1 U4I
*> 4-

41 n 0 0 4- 0

m Co 0 0. 0
4 $ f V 1 V V
4M 0 4 0 4 v 4 41 0 04
lz U O- 0 O4
V0 0 W V
2 U 4.4 .. 414 41 441 V P

0 w w 0H w 04.' w 0 V)
C0 3 4 4-i ,a V 4-4

0) 0 D O 0 p 4-4 p
w0 a) -i 4u 1 < 1

S c 41 0 0 41 Q

41 3 4 r P 4 0 U e4 0B

,> 06p 4 0 *cr ,4

(0 0 CU 44 0 -H n t -l

6 4 a
'I. NV


N -i


S- *- 0
-. 4- .M -

*0 *

%4 a 9%.
4 0-xxx

>, B *J t 0 0<"
bL ** bb ->
*o ^ a '" &*o.
* u **<

supplementation, improved feed intakes by beef cattle and sheep (Reid,

1983; Reid, 1979). Supplying MgO in a grain-mineral mixture raised

dietary Mg concentrations from .16 to .45% of dry matter and increased

weight gain of pregnant beef cows, and stover intake by dry beef cows in

balance trials (Reid, 1983). Other research showed omitting Mg from the

diet of sheep caused intakes to decline to 32% of preliminary period

intakes after 4 d of feeding Mg-deficient diets (Ammerman et al.,

1971). Similar results were seen in a study with lambs, yearlings and

adult sheep in which voluntary intakes were reduced to 76, 59 and 58% of

preliminary period intakes after 3 d of feeding Mg-deficient diets

(Chicco et al., 1973a). There were no significant main effects of

dietary Na or K on feed intake in our experiment. Schneider et al.

(1986) found lactating cows consumed 4.5% more dry matter as dietary K

increased from 1.3 to 1.8%. However, in another experiment 1.5% total K

did not improve feed intakes compared with diets containing 1.0% dietary

K (Schneider et al., 1984). Feeding .55% total dietary Na (.85%

NaHCO3, supplemented) improved feed intakes of lactating cows 7.2%

(Schneider et al., 1984).

Actual milk yields (not adjusted for milk fat percent or dry matter

intake) responded in curvilinear fashion to added dietary Mg (P<.01)

(Table 5-4). Actual milk yields increased 5.2% as dietary Mg increased

from .26 to .48%. Four percent fat-corrected milk yields (4% FCM)

(Tyrrell and Reid, 1965) increased as dietary Mg increased from .26 to

.48% (7% increase) and then declined with .60% Mg (overall curvilinear

response, P<.01). Identical responses to dietary Mg for actual milk

yields and 4% FCM yields were seen when dry matter intake was included

as a continuous independent variable in the analysis of variance (data

not shown). Wilson (1980) reported an 8.5% increase in milk yields when

dairy cows were drenched daily with 10 g of Mg supplied as magnesium

chloride. Milk yield responses in our study were similar to those

reported by Teh et al. (1985). In their study lactating cows fed a

50:50, concentrate:corn silage diet (dry basis) were supplemented with

0, .40 or .80% MgO giving corresponding total dietary Mg contents of

.23, .44, and .57% Mg. Actual daily milk yields were 9.8% greater with

.44% Mg than with .23% Mg. The actual milk yield responses reported by

Teh et al. (1985) did not appear to be caused by alkalizing and (or)

buffering effects of MgO since there were no differences in ruminal pH

or ruminal acetate:propionate ratios among cows fed 0, .40, or .80%

dietary MgO. In other studies with MgO supplementation to lactating

cows, milk yield responses were not found (Thomas and Emery, 1969;

Erdman et al., 1980; Jesse et al., 1981; Erdman et al., 1982; Thomas et

al., 1984), possibly due to excessively high dietary Mg (greater than

.6% Mg) which may have caused adverse effects on ruminal fermentation

and (or) dry matter intakes. There were no effects of dietary Na or K

on actual or 4% FCM yields. This contrasted with Schneider et al.

(1984) who found increased milk yields by lactating cows during the

summer when dietary K increased from 1.0 to 1.5% and NaHCO3 was added

at .85%. In another experiment, milk yields (with dry matter intake in

the model as a continuous independent variable) were improved 4.2% by

1.0% NaHCO3 and 5.3% by .73% NaC1 (Schneider et al., 1986). Actual

milk yield (not adjusted to equal dry matter intake) was improved 4.2%

by increasing dietary K from 1.3 to 1.8% (Schneider et al., 1986).

In our study milk fat percentage was unaffected by added dietary Mg,

Na, or K. We were not surprised that we were unable to find any effect

of Mg on milk fat percentage because the basal diet was not

fat-depressing (mean milk fat percentage was 3.59). Similar results

were found by Teh et al. (1985) where milk fat percentage was not

improved by .40 or .80% MgO (mean milk fat percentage in their study was

3.64). Wilson (1980) found drenching dairy cattle with 10 g Mg daily

did not improve milk fat percentage although milk yields increased

8.5%. Most effects of added dietary MgO to raise milk fat percentages

were observed when milk fat-depressing diets were fed (Thomas and Emery,

1969; Jesse et al., 1981; Teh et al., 1985). In each of these studies

results suggested milk fat percentages were improved by an alkalizing

and (or) buffering effect of MgO in the digestive tract. However, it is

a common physiological response that milk fat content increases as milk

yields decline. Thus, in these experiments it was difficult to

differentiate between increases in milk fat percentages due to decreased

milk yields or ruminal alkalizing and (or) buffering effects of dietary

MgO. In the current experiment milk fat yields increased from .87 kg/d

to .94 kg/d as dietary Mg increased from .26 to .48% and then declined

to .90 kg/d with .60% Mg (curvilinear effect, P<.01). Increased fat

yield was a result of increased milk yields (P<.01). Others found MgO

in diets for lactating cows increased fat yields due to increased milk

fat percentages with no corresponding changes in milk yields (Jesse et

al., 1981; Thomas et al., 1984). Milk protein percentage declined

linearly from 3.31 to 3.27 as dietary Mg increased from .26 to .60%

(P<.05). Decreased milk protein percentages found in our study agreed

with results of Thomas et al. (1984) where cows fed either .50% MgO (.33

to .72% total dietary Mg), .55 to .70% Mg(OH)2 (.60% total dietary

Mg), or 1.0% NaHCO3 (.20% total dietary Mg) had significantly lower

milk protein percentage than cows not receiving any mineral

supplements. Other studies showed no effects of MgO on milk protein

percentage or daily milk protein output (Emery and Brown, 1961; Miller

et al., 1965; Thomas and Emery, 1969; Kilmer et al., 1980; Erdman et

al., 1982). Schneider et al. (1986) reported milk protein percentage

declined slightly as dietary K increased from 1.3 to 1.8%. In our study

milk protein yields increased from .83 to .91 kg/d as dietary Mg

increased from .26 to .48% and then declined with .60% Mg (curvilinear

effect, P<.001). Added Na decreased milk protein yields 4.7% (P<.05).

Milk and Plasma Mineral Concentrations and Fractional Excretions

Least squares analysis of variance and least squares means for milk

mineral composition are in Tables 5-5 and 5-6. Overall, concentrations

of Mg, Ca, K, Na and P in milk were similar to reported values (Kemp et

al., 1961; Rook et al., 1964; Kemp and Geurink, 1978; Mallonee, 1984;

Schneider et al., 1984; Schneider et al., 1986). Dietary Mg had no

effects on milk Mg or Na concentrations. There was a trend towards

dietary Mg affecting milk K and P in a cubic fashion (P<.10). There

also tended to be a linear effect of Mg on milk Ca concentrations,

decreasing from 1336 ppm at .26% Mg to 1247 ppm at .60% Mg (6.7%

decrease) (P<.11). There were no detectable main effects of dietary Na

or K on mineral concentrations in milk. Schneider et al. (1984)

reported a 4.6% increase in milk K when dietary K was increased from 1.0

to 1.5%. However, in another study (Schneider et al., 1986) milk K

concentrations were unchanged by increasing dietary K from 1.3 to 1.8%.

TABLE 5-5.


Mean squares

Source df Milk Mg Milk Cab Milk Kb Milk Nab Milk pb

------------------------ (ppm) -------------------------
Period 2 1293.57*** 348.19*** 34.48* 5.69 765.15***
Cow 46 405.50*** 45.25* 42.59*** 22.30*** 47.80
Na 1 7.49 2.98 3.68 <.01 6.43
K 1 140.35 12.20 2.56 4.34 20.93
MgL i 48.90 60.25 1.18 6.88 46.08
MgQ 1 11.55 4.92 1.33 2.35 24.76
MgC 1 116.01 47.70 28.84+ 1.78 101.92+
Na X K 1 91.19 46.70 6.77 .03 2.44
Na X MgL 1 24.03 5.46 13.00 .47 2.26
Na X MgQ 1 54.92 1.71 5.09 .34 2.09
Na X MgC 1 50.43 <.01 5.72 3.64 3.76
K X MgL 1 30.18 1.73 .27 .92 40.52
K X MgQ 1 40.17 12.22 17.62 .40 38.38
K X MgC 1 56.91 2.56 4.50 .91 7.42
Na X K X MgL 1 53.56 5.82 32.67+ 5.36 .24
Na X K X MgQ 1 1.46 48.70 1.58 .12 56.00
Na X K X MgC 1 29.33 <.01 1.16 .35 13.04
Residual 74 94.38 22.78 8.93 3.22 36.07

aNa sodium, K potassium, Mg
MgL = linear magnesium effect,
cubic magnesium effect.

- magnesium,

Ca calcium, P phosphorus,

MgQ quadratic magnesium effect, MgC =

bMean squares for Ca, K, Na and P are tabular values times 103

Error term for all sources of variation.






Dietary main effectsab

Mineral .26%

Dietary Mg
.38% .48%

Dietary K
.60% 1.14% 1.59%

Dietary Na
.24% .62%

Ca 1336 1253 1304 1247 1295 1275 1279 1291
Mg 136 132 135 133 135 133 134 134
K 1670 1658 1718 1690 1687 1696 1685 1698
Na 468 442 449 441 456 444 450 450
P 1419 1358 1445 1326 1400 1374 1396 1378

aLeast squares means.

bMg magnesium, Na = sodium, K potassium, Ca -

calcium, P =

CCubic Mg effect (P<.10).

Least squares analysis of variance and least squares means for

plasma concentrations and renal fractional excretions of minerals are in

Tables 5-7 and 5-8. As dietary Mg increased from .26 to .60%, plasma Mg

concentrations increased linearly from 2.52 mg/100 ml to 2.68 mg/100 ml

(P<.01). We have reported increases in plasma Mg concentrations of cows

administered Mg acetate intraruminally (Chapter 3). Others noted

higher dietary Mg increased plasma or serum Mg concentrations (Erdman et

al., 1982; Thomas et al., 1984; Teh et al., 1985). Plasma K

concentrations were increased 4.0% by higher Na in our study (P<.01).

There were no effects of K on plasma K concentrations in agreement with

Schneider et al. (1984; 1986). Similarly, there were no effects of

dietary Na on plasma Na concentrations, in agreement with Schneider et

al. (1984; 1986). There was a trend towards plasma P concentrations

responding in a curvilinear fashion as dietary Mg increased from .26 to

.60% (P<.10).

Fractional excretions were calculated as the ratio of urine mineral

concentration to plasma mineral concentration divided by the ratio of

plasma creatinine to urine creatinine concentrations (Lane and Merritt,

1983). Fractional excretion (FE) of a specific mineral is an estimate

of the percentage of the total mineral load filtered through the kidney

which is excreted. Least squares means for fractional excretions of Mg

(FEMg) in our study ranged from 20.3 to 30.0%, similar to values of

Thomas et al. (1984); their values ranged from 13.6 to 30.8%. Added

dietary Mg tended to increase FEMg curvilinearly from 21.0% with .26% Mg

to 30.0% with .48% Mg (42.9% increase), then declining to 21.2% with

.60% Mg (P<.16). Thomas et al. (1984) found that adding .50% MgO or











4IN 0o o me-0 N400u4m


I o0s..-. 0 441 OOu m0'o0N-.

0, 0! 000 0000 -


In m o-c0o00%- 0 0o0
I Y- 0-0000000o-00000

f .41

IC ... N0. ...0 .......

00 000000-000000-01

I +
1 41 4. CIO col h4 CP Chr >o r,

C- -Ne t4 -4
10 0 0 0M M Mt"O 0 o
1- N I0V4 1' 0 .0

1N fl <0 CN N

1, 4 t 4. 4.
004- n N-o- -004N-10-.1
S' * V

IoN 100m.01-4-0 0 4- -.sw No
'04N NU4 -ON n

I --^ N

I... **

IN -" N Ne a X O

a 6ve kl I
*.- c uuk~u


0r 4

o u


bo 4-
B *

i-I 0)

bO -u

u o0
-H -4 "4
4. U 0 0
W 4 0) 0
C C Co M Or
60 0 01U

(o 0 0 w0

a 4a 0

0c 4) 0 0
4J4 4 -0 (B (0


Co 0 to *4
0 4 c tf r 4 -

Sm C M 44
0o r0 30

C0 0 C CO 0 V
o 0 0 0 1 4 V o
M -4 C0 CW V *
i-40 4 ( -IC C V

(i r D 0 + Q) I C-K -I

oo 0 N 00

N %
SO CM) 0 0G

n 0% N t-: 0

N 0%
S00 0 00

N 0


C 000 0.

CN 0 N0 C- 0

CN 0%

N 0000

*A 0--O o

- 00 0
CM0" C
-i CM cy





44 z


H f


0 0



s H





e *





4) 0
1 <

0 C 000

4 %C-


N 0

NM a0 -:
0 O ~000

00 00
4 -- 9

- 4'-N0%

4 '0o -4

0 ON0M
N '0

No Ios -












0 -

Do V-V V 0
03 0 4 4

0 P + 0 PV 0 P4

k0 Os 44 0 ) "(d *44 0 4.) k
. Q 0 10 ) 44
S 0 d to 4o a)

&0 CO O & 0 H

v U 44U -bD X ll d

-l -ua

.55% to .70% Mg(OH)2 to the diet increased FEMg approximately 68.0%,

compared with no supplemental Mg. Fractional excretions of Ca (FECa) in

our study, .09 to .92%, were in the same range as those of Thomas et al.

(1984) which ranged from .47 to 2.00%. In our study, added Mg depressed

FECa linearly (P<.01) (Table 5-8). The FECa response tended to be

curvilinear (P<.05). Thomas et al. (1984) found FECa was depressed

53.0% in lactating cows fed .50% MgO or .55 to .70% Mg(OH)2.

Fractional excretions of Ca increased (106.7%) as dietary Na increased

from .24 to .62% (P<.10). Fractional excretions of Ca were decreased

78.9% as dietary K increased from 1.14 to 1.58% (P<.001). In other

experiments a similar effect of K was reported (Deetz et al., 1982;

Greene et al., 1983). Main effect least squares means for fractional

excretions of K (FEK) ranged from 43.3 to 109.3%. Fractional excretion

with 1.14% dietary K was 51.2%, similar to other reports (Mallonee,

1984; Thomas et al., 1984). In our study FEK increased from 51.2 to

81.4% by elevating dietary K from 1.14 to 1.59% (P<.001). Fractional

excretions of Na (FENa) were low, ranging between 1.8 and 3.3%, similar

to results of Mallonee (1984). This is typical of the ruminant which

conserves Na very tenaciously. As dietary Na increased from .24 to

.62%, FENa increased 73.7%. Thomas et al. (1984) found 1.0% added

NaRCO3 increased FENa 280.0% compared with no added NaHCO3. Least

squares means for fractional excretions of P ranged from 9.0 to 13.7%.

These values were higher than those of Thomas et al. (1984), although

urinary concentrations of P in our study ranged between 18.7 and 43.3

mg/100 ml similar to Thomas et al. (1984). Fractional excretions of P

increased 35.6% as dietary K increased from 1.14 to 1.59% (P<.05).

Fractional excretions of P tended to decrease linearly with increasing

dietary Mg (P<.10).

Results suggested that dietary Mg requirements for lactating dairy

cows are higher than those stipulated by the NRC (1978). When dry

matter intake was included in the mathematical model as a continuous

independent variable (covariate), significant production responses to

dietary Mg remained, suggesting that additional Mg enhanced digestion

and nutrient assimilation and (or) milk synthesis at the mammary gland.

We do not have digestibility data from this experiment. It does not

appear that lactational reponses to Mg reported here resulted from

effects of MgP on ruminal alkalizing and (or) buffering actions thus

improving digestion. However, in a companion laboratory experiment we

compared the alkalizing and (or) buffering potential of two MgO sources,

two levels of the feed grade MgP used in our lactation experiment and

control (no Mg source addition). Treatment additions of 1 g of either

MgO-A (52% Mg, Magal, Magnesitas de Rubian, Madrid, Spain), MgO-B (58%

Mg, Southeastern Minerals, Chamblee, GA) 1.0 g MgP, or 2.42 g MgP were

weighed into 150 ml breakers. The 2.42 g MgP treatment was included

because we were interested in the alkalizing and (or) buffering

capacities of MgP compared with 1 g MgO if approximately equal Mg were

contributed from each source. Within each of four replications of the

experiment the procedure initially was to add 50 ml of distilled water

to beakers containing the Mg sources. Contents were hand-swirled for 5

min. Individual beakers were selected randomly (within replication) and

stirred on a magnetic stir plate for 2 min. After 2 min the beaker was

removed from the stir plate and pH electrodes were placed in the

suspension. After exactly 2 min standing time, pH was recorded. After

pH of each beaker within replication was recorded, 5 ml of .1024 N

sulfuric acid was added to each and the same procedure for pH

determination was repeated; 5 min with hand-swirling, 2 min stirring, 2

min standing time and recording of pH. This procedure was repeated with

sequential 5 ml additions of sulfuric acid until a total of 40 ml had

been added. After the acid addition phase of the experiment all beakers

(four replications) were saved and the final pH was determined after

beakers stood for 20 h. Data were analyzed by method of least squares

analysis of variance using general linear model procedures of SAS

(1982). Mathematical model and degrees of freedom (df) included:

replication (3 df), magnesium treatment (4 df), sequential acid addition

(8 df), replication by treatment interaction (12 df), replication by

acid addition interaction (24 df), treatment by acid addition

interaction (32 df) and error (96 df). Orthogonal contrasts were used

to compare overall treatment means and pH changes with successive acid

additions among treatments. Overall, the pH of the control (water)

treatment (pooled across acid additions) was lower than Mg source

treatments (P<.001). Suspensions with MgO sources had higher pH than

MgP treatments (P<.001). Comparing MgO sources, MgO-A had a higher

overall pH than MgO-B (P<.001). Suspensions of either MgO source raised

pH to 10.677 and to 11.204 compared with distilled water (no sulfuric

acid addition, pH = 7.974) (Figure 5-1). The suspensions containing

MgO-A and MgO-B resisted pH change with sequential additions of acid.

After total acid additions of 40 ml, pH of the MgO-A suspension was

9.569 compared with 9.149 for the MgO-B suspension. In contrast, 1 g or