@Copyright 1999 by Humana Press lnc.

All rights of any nature, whatsoever,reserved.

01634984/99/6702-0173 $13.50

Silicon Bioavailability Studies

in Young Rapidly Growing Rats
and Turkeys Fed Sernipurified Diets
A Comparative Study
HENRY K A Y O N G O - M A L E *'1 AND XIUJUAN J I A 2

'Departments of Biology and Microbiology and 2Chemistry

and Biochemistry, South Dakota State University, Brookings,
South Dakota 57007-0595
Received J a n u a r y 10, 1998; Revised April 3, 1998;
Accepted April 15, 1998

ABSTRACT
Two experiments were conducted using completely randomized
designs to study the bioavailability of Si from three sources to growing rats and turkeys fed semipurified diets. The basal diets were
dextrose-egg albumin for rats and dextrose-casein for turkeys. The Si
sources were tetraethylorthosilicate (TES), sodium silicate (NaSil), and
sodium zeolite A (NaZA). Rats and turkeys were supplemented at 500
and 270 ppm Si, respectively, from each source. A control group of
unsupplemented rats and turkeys was included in each experiment.
In general, irrespective of Si source, Si supplementation slowed
(p < 0.05 or p < 0.01) growth rates in both rats and turkeys. Although
dietary Si supplementation reduced (p < 0.05) plasma Mg levels and
liver Zn concentrations in rats, it increased (p < 0.05) plasma P and
reduced (p < 0.05) plasma Cu levels in turkeys.
Rats on TES had significantly slower (p < 0.05 or p < 0.01) growth
rates (5-10~ than those on NaSil or NaZA. In rats, NaZA and TES
reduced (p < 0.05) hemoglobin concentrations and plasma Zn, respectively. However, plasma Mg levels were higher (p < 0.05) in TES than
NaSil- or NaZA-fed rats. The source of the dietary Si did not affect
(p < 0.05) the organ weights of rats and their mineral concentrations.
Turkeys on TES diets grew at a significantly faster (p < 0.05) rate
(15%) than those on NaSil or NaZA diets during the first 2 wk of
experimentation. However, after 4 wk, there were no significant
*Author to whom all correspondence and reprint requests should be addressed.
Biological Trace Element Research

] 73

Vol 67, 1999

Kayongo-Male and Jia

174

(p > 0.05) differences in growth between the Si sources. In turkeys,

NaZA increased (p < 0.05) hematocrit levels and plasma Mg levels.
Turkeys on NaZA diets had larger (p < 0.05) hearts and livers than
those on NaSil but not TES. Liver Mn content was higher (p < 0.05) in
turkeys on NaSil than TES or NaZA. Heart Zn was lower (p < 0.05)
in turkeys on NaSil than TES, but not NaZA.
Index Entries: Silicon, sodium silicate, sodium zeolite, tetraethylorthosilicate, tissue minerals, serum alkaline phosphatase.

INTRODUCTION
Silicon is the second most abundant element on earth (1). A large
amount of silicon occurring in nature is in the form of aluminosilicates and
polymeric silica and is, therefore, not biologically readily available (2). Animal experiments have shown that the bioavailability of ingested silicon is
very small and that the form of dietary silicon is an important factor determining the rate of production of soluble or absorbable silicon from various
silicon compounds in the gastrointestinal tract (3,4). Some forms of silicon in
foods and beverages are readily absorbed from the gastrointestinal tract (2).
Tetraethylorthosilicate and sodium silicate have been reported to
yield readily absorbable monosilicic acid in gastric stomachs with p H
levels of less than 3.0 (5-7). Monosilicic acid is freely diffusible in tissue
fluids (1,3). Sodium metasilicate, like tetraethylorthosilicate, has been
used widely by researchers to study the biological functions of silicon
(8,9). Sodium zeolite A containing about 15% silicon has been used as a
silicon source by Wiegand et al. (10) to study silicon bioavailability in
rats. Sodium zeolite A has a high silicon bioavailability (10). The drawback with sodium zeolite A is that it also contains 13% aluminum, which
may be toxic to some biological systems. Aluminum displaces magnesium
and calcium from active sites of key enzymes, thereby inhibiting the catalytic action of those enzymes (11). Aluminum also disrupts calciumdependent electrophysiological functions of muscles and may cause
mortality from cardiovascular and respiratory dysfunctions (12).
The objective of this study was to compare the bioavailability of silicon from three sources: sodium silicate, tetraethylorthosilicate, and
sodium zeolite A to young, rapidly growing rats and turkeys fed semipurified, dextrose-albumin- or dextrose-casein-based diets.

MATERIALS AND METHODS

Two experiments were conducted with young, rapidly growing rats
and turkeys, using semipurified based diets. The basal diets were dextroseegg albumin for rats and dextrose-casein for turkeys. The rapid growth
phases of rats and turkeys are 8-12 and 4-6 wk of age, respectively. Three
silicon compounds were used as sources of silicon (Si) to provide 270- and
Biological Trace Element Research

Vol. 67, 1999

Silicon Bioavailability Studies

175

500-ppm Si levels of dietary inclusion for turkeys and rats, respectively.

The three sources were (1) tetraethylorthosilicate or TES (J. T. Baker
Chemical Co., Phillisburg, NJ) which is readily hydrolyzed in gastric
stomach (pH < 3.0) to monosilicic acid that is available for absorption,
(2) sodium silicate (Na2SiO3 95H20) or NaSil (Matheson, Coleman and Bell,
Norwood-Cincinnati, OH), and (3) sodium zeolite A or NaZA (EthacalTM,
Ethyl Corp., Baton Rouge, LA).

Experiment 1
Seventy-two (72) male Sprague-Dawley albino rats (SASCO, Omaha,
NE) initially weighing, on average, 45.0 g were allotted into a fourdietary-treatment completely randomized experiment. The basal diet
(Table 1) was a dextrose-egg-albumin type. In order to reduce avidin
interference with biotin absorption, the spray-dried egg albumin was
denatured by autoclaving at 120~ for 1 h and drying at 110~ for 1 h.
The four dietary treatments included a control diet that inherently contained < 5.0 ppm Si, and three other diets supplemented with 500 ppm
Si each obtained by additions of TES or NaSil or NaZA.
Animal care and management was similar to that described by
Emerick and Kayongo-Male (13). Rats were weighed weekly during the
8-wk experimental period. At termination of the experiment, blood was
drawn via heart puncture into heparinized vacuum tubes. Rats were
killed by decapitation while under anesthesia. The heart, liver and
femurs were collected from each rat and processed for chemical analyses
as described by Emerick and Kayongo-Male (13).

Experiment 2
Seventy-two (72) newly hatched Nicholas tom turkey poults (Willmar
Poultry Co., Willmar, MN), weighing on average 54.3 g initially, were allotted into a four-dietary treatment, completely randomized experiment. The
poults were caged in stainless-steel cages with elevated wire-mesh floors
and were housed in the animal room having controlled temperature and
lighting. The room temperature was kept at 35~ 32~ 29~ and 24~ during the first, second, third, and fourth week of the experiment, respectively.
The basal diet (Table 2) was a 28% protein, dextrose-casein type formulated to mimic the recommended nutrient content of the turkey starter
diet (14). The four dietary treatments included a control diet (0 ppm Si)
and three other diets supplemented with 270 ppm Si each, obtained by
additions of TES or NaSil or NaZA.
The birds were weighed weekly during the 4-wk experimental
period. At termination, a blood sample was drawn from each turkey via
the V. provunda branchii wing vein into heparinized vacuum tubes. The
birds were then killed by cervical dislocation followed by exsanguination.
The heart, liver, and tibia were removed from each bird and processed for
chemical analyses as described by Emerick and Kayongo-Male (13).
Biological Trace Element Research

Vol. 67, 1999

176

Kayongo-Male a n d Jia

Table 1
Rat Basal Diet Compositiona
Ingredient

Amount (%)

Dextrose, anhydrous

70.5

Egg albumin, spray dried

20.0

Salt mixture b

3.5

Vitamin mixture~

1.0

Corn oil

5.0

aDiet composition before the additions of the Si sources, as dietary

variables.
bThe percentage composition of the salt mixture (AIN-76 salt mixture, ICN Biochemicals,Inc., Cleveland, OH, USA): calcium phosphate
dibasic, 50.0; sodium chloride, 7.4; potassium citrate monohydrate, 22.0;
potassium sulfate, 5.2; magnesium oxide, 2.4; manganous carbonate
(43-48% Mn), 0.35; ferric citrate (16-17% Fe), 0.6; zinc carbonate, 0.66;
cupric carbonate (53-55% Cu), 0.030; potassium iodate, 0.001; sodium
selenite, 0.001; chromium potassium sulfate, 0.055.
cAIN-76 vitamin mixture (ICN Biomedicals, Inc) composition in
grams/kilogram mixture triturated in dextrose: thiamine hydrochloride, 0.6; riboflavin, 0.6; pyridoxine hydrochloride, 0.7; nicotinic acid
3.0; D-calcium pantothenate, 1.6; folic acid, 0.2; D-biotin, 0.02; vitamin
(cyanochobalamin)B12,0.001; vitamin A (retinyl palmitate, 250,000 IU/g),
1.6; vitamin E (DL-alpha-tocopherolacetate, 250 IU/g), 20; vitamin D3
(cholecalciferol,0.25; vitamin K (medaquinone), 0.005.

General Procedures
In both experiments, a 12-h day-and-night cycle was maintained. An
attempt was m a d e to minimize silica contamination by initially cleaning
the animal room and related equipment followed by a weekly flushing
of the painted concrete floor. Every day, the birds and rats were given
ad libitum access to fresh diets and distilled water provided in stainlesssteel feeders and water containers.
Chemical Analyses

Determination of hematocrit and hemoglobin were done immediately following the collection of the blood samples using standard procedures described by Emerick and Kayongo-Male (13). Plasma
separated from heparinized blood samples and stored at -25~ was
used for cholesterol, enzyme, and mineral assays. Plasma total cholesterol was d e t e r m i n e d using enzymatic kits (Sigma kit 352-50, Sigma
Chemical Co., St. Louis, MO). Serum alkaline phosphatase activity was
determined by the hydrolysis of p-nitrophenyl phosphate (Sigma kit 104,
Sigma Chemical Co., St. Louis, MO).
Biological Trace Element Research

Vol. 67, 1999

Silicon Bioavailability Studies

177

Table 2
Turkey Basal Diet Compositiona
Ingredient

Amount (%)

Dextrose, anhydrous

51.2

Casein

33.0

Solka flock

3.0

Corn oil

4.0

L-Arginine

0.4

DL-Methionine

0.4

Mineral mixture~

7.0

Vitamin mixture~

1.0

aDiet composition before the additions of the Si sources, as

dietary variables.
bMineral mixture composition in grams/kilogram triturated in
finely powdered sucrose: calcium diphosphate, dibasic, 313; calcium
carbonate, 197; sodium chloride, 72.6; potassium chloride, 21.0; potassium citrate, monohydrate, 246; magnesium sulfate, anhydrous, 42.4;
manganous carbonate, 1.8; ferric citrate, 6.71; zinc carbonate; 1.92;
cupric carbonate, 0.024; chromium potassium sulfate. 12H20, 0.22; tin
sulfate, 0.09; potassium iodate, 0.01; sodium selenite, 0.01.
cVitamin mixture composition in grams/kilogram triturated in finely
powdered sucrose: vitamin A palmitate (500,000 IU/g), 1.0; vitamin D3
(400,000 IU/g), 0.30; vitamin E acetate (500 IU/g), 4.0; menadione
sodium bisulfite, 3.0; choline bitartrate, 211; thiamine HC1, 0.22;
riboflavin, 0A; pyridoxine HC1, 0.55; niacin, 6.00; calcium pantothenate, 1.10; folic acid 0.1; biotin, 0.03; vitamin B12 (0.1%), 1.0.

Calcium (Ca), magnesium (Mg), and zinc (Zn) in plasma plus copper (Cu) and Zn in nitric acid digests of liver and heart tissues were
determined by flame atomic absorption spectrophotometry (PerkinElmer Model 5100PC, Norwalk, CT) as described by Stewart et al. (15).
Phosphorus (P) in plasma was determined using the Fiske and Subbarow
phosphomolybdate method (16). Plasma Cu was assayed by graphite furnace atomic absorption (Model 5000 equipped with Model 500 heated
graphite atomizer and autosampler, Perkin-Elmer) using 0.2% nitric acid
(Utrex, J. T. Baker Chemical) as diluent.
Statistical analysis of the data was done by the general linear model
(GLM) analysis of variance (ANOVA) procedures using a model consisting of four dietary treatments based on a Si source for the two experiments (17). The Fisher-protected least square difference (LSD) test
Biological Trace Element Research

Vol. 67, 1999

178

Kayongo-Male and Jia

was used to identify significant differences among treatment means.

Standard errors of the means were calculated from the error mean
squares (18).

RESULTS
Growth Data

Rats
The rat growth data are shown in Fig. 1. The addition of dietary Si
from the different sources significantly affected (p < 0.05 or p < 0.01) rat
body-weight changes. From the third to the fifth week of Si supplementation, the rats on TES had a significantly slower (p < 0.05) growth rate
than those on NaSil but not NaZA. During the sixth week, the rats on TES
had a significantly slower (p < 0.05) growth rate than those on NaZA but
not NaSil. During the seventh and eighth week, rats on TES had a significantly slower (p < 0.05) growth rate than those on NaSil or NaZA. After
8 w k of Si supplementation, rats on TES, NaSil and NaZA weighed, on
average, 235, 257, and 268 g, respectively. At the same time, the weight of
the rats on the control diet was 273 g, which was significantly higher
(p < 0.05) than that of rats on the TES but not on the NaSil or NaZA diets.

Turkeys
The turkey growth data are shown in Fig. 2. Dietary supplemental Si
from different sources significantly affected (p < 0.05 or P < 0.01) turkey
growth. During the first week of silicon supplementation, the turkeys on
NaSil diet had a slower (p < 0.05) growth rate than turkeys on the control
and TES diets but not those on the NaZA diet. During the second week, the
control and TES turkeys grew faster (p < 0.05) than turkeys on NaSil or
NaZA. However, in the third week, turkeys on the control diet grew faster
(p < 0.05) than those on NaSil or NaZA diets but not those on TES diets.
Within the Si-supplemented groups, turkeys on TES diets grew faster
(p < 0.05) than those on NaSil or NaZA diets during the second week but
not during the third week of the experiment. After 4 wk of Si supplementation, there were no significant differences (p > 0.05) in the growth rates
of all turkeys on the four dietary treatments although the unsupplemented
turkeys weighed more than the Si-supplemented turkeys (834 vs 760 g).

Blood Data

Rats
The rat blood data are presented in Table 3. Hemoglobin content was
significantly affected (p < 0.05) by Si source. Rats on the NaZA diet had
lower (p < 0.05) hemoglobin levels than rats on the rest of the diets.
Plasma Ca, Mg, and Zn were significantly affected (p < 0.05) by Si source
but not plasma Cu and P. Plasma Ca content was lower (p < 0.05) in rats
Biological Trace Element Research

Vol. 67, 1999

Silicon Bioavailability Studies

179

Fig. 1. Body-weight changes of rats fed diets containing silicon from various
sources during the 8-wk experimental period. Silicon sources: Cont = control with
no silicon added; NaSil = sodium silicate (Na2SiO3.5H20); TES = tetraethylorthosilicate; NaZA = sodium zeolite A (15% silicon); SEM = standard error of the
mean calculated from the error mean squares; a, b, c = graph columns, within 1 wk
on the experiment, not sharing a common letter(s) differ at p < 0.05.
on NaSil diet t h a n the control diet b u t not the TES or N a Z A diets. Plasma
M g levels w e r e h i g h e r (p < 0.05) in the control rats t h a n the Si supplem e n t e d rats. Within the Si s u p p l e m e n t e d rats, rats on the TES diet h a d
h i g h e r (p < 0.05) M g levels t h a n those on NaSil or N a Z A diets. Rats on
TES diets h a d lower (p < 0.05) p l a s m a Z n levels t h a n those on the control a n d N a Z A diets b u t not NaSil diets.

Turkeys
The t u r k e y blood data are s h o w n in Table 4. H e m a t o c r i t levels a n d
not h e m o g l o b i n content w e r e significantly affected (p < 0.05) b y the Si
Biological Trace Element Research

Vol. 67, 1999

180

Kayongo-Male and Jia

Fig. 2. Body-weight changes of turkeys fed diets containing silicon from various sources during the 4-wk experimental period. Silicon sources: Cont = control
with no silicon added; NaSil = sodium silicate (Na2SiO3.5H20); TES = tetraethylorthosilicate; NaZA = sodium zeolite A (15% silicon); SEM -- standard error of the
mean calculated from the error mean squares; a, b, c = graph columns, within I wk
on the experiment, not sharing a common letter(s) differ at p < 0.05.

source. Turkeys on the N a Z A diet h a d higher (p < 0.05) hematocrit levels than turkeys on the control diet b u t not those on the NaSil or N a Z A
diets. Plasma P, Mg, and Cu were significantly affected (p < 0.05) by the
Si source b u t not p l a s m a Ca and Zn. Turkeys on the control diet h a d
lower (p < 0.05) p l a s m a P levels than the s u p p l e m e n t e d turkeys w h o s e
P levels were not different (p > 0.05) b e t w e e n the Si sources. Plasma Mg
levels were higher (p < 0.05) in the turkeys on the N a Z A diet than the
t u r k e y s on the control a n d NaSil diets b u t n o t TES diet. Turkeys on
the control diet h a d h i g h e r (p < 0.05) p l a s m a Cu levels t h a n the Sis u p p l e m e n t e d turkeys w h o s e Cu levels were not different (p > 0.05)

Biological Trace Element Research

Vol. 67, 1999

Silicon Bioavailability Studies

181

,,6

,..4

~.

c~

r~

"~

c~

c~

c~

t~

~2:

~1

- <

Biological Trace Element Research

Vol. 67, 1999

182

Kayongo-Male a n d Jia

II

~I

I~.

~d

,..

#.
>

"~,

~a
-a
9
9

~.~

~i ~ ~
Biological Trace Element Research

<

Vol. 67, 1999

Silicon Bioavailability Studies

183

Table 5
Mineral Concentrations and Weights
Mineral Concentrations
Dietary
Treatment

Heart
Copper

Weights ~

Liver
Zinc

..........

Copper

Heart

Liver

Zinc

ppm, dry basis . . . . . . . . . . . . . . .

% ......

Control

23.3

79.7

12.8

95.2 a

0.33

3.67

TES

23.2

78.7

12.5

89.1 b

0.37

3.96

NaSil

22.1

76.0

11.7

88.5 b

0.35

3.74

NaZA

21.6

77.4

11.5

88.9 b

0.33

3.66

1.76

0.01

0.10

NS

NS

SEM2

1.11

1.94

GLM3

NS

NS

0.44
NS

.025

1Heart and liver weights expressed as percent of final (8th week) 'bodyweight'
Standard error of the mean calculated from the error mean square.
A one-way analysis of variance using Type III sum of squares; numerical values are
levels of probability; NS = not significant (p > 0.05)
Least square means within a column not sharing a c o m m o n superscript letter differ
at p < 0.05.

between the Si sources. Total plasma cholesterol and serum alkaline

phosphatase activity were not affected (p > 0.05) by the Si source.

Organ Data
Rats
The rat liver and heart weights and their mineral concentrations are
shown in Table 5. The source of Si did not affect (p > 0.05) organ weights
or their mineral concentrations except liver Zn concentrations. Liver Zn
concentrations were higher (p < 0.05) in the unsupplemented than the Sisupplemented rats. Within the Si-supplemented groups, liver Zn levels
did not significantly (p > 0.05) differ.

Turkeys
The turkey liver and heart weights and their mineral concentrations
are shown in Table 6. The source of Si significantly (p < 0.05) affected
heart and liver weights and liver Mn and heart Zn concentrations. Turkeys on NaZA diets had larger (p < 0.05) hearts and livers than turkeys
on the control and NaSil diets but not those on the TES diets. Turkeys supplemented with NaSil had the smallest hearts and livers. Liver Mn content was higher (p < 0.05) in turkeys on NaSil than TES and NaZA diets
but not the control birds. Heart Zn was lower (p < 0.05) in turkeys on the
control and NaSil than TES diets but not the NaZA diets.
Biological Trace Element Research

Vol. 67, 1999

Kayongo-Male and Jia

184

II

~1~

t~

"~"

0'~

9
V

&

"~_~=

~oo

il
~.~ ~

'-~

.g

~-~.~
"S_>, . ~

.~

or~
9

Biological Trace Element Research

Vol. 67, 1999

Silicon Bioavailability Studies

185

DISCUSSION
The NaSil- and NaZA-supplemented diets supported growth much
better than TES in rats. The differences in growth rates of rats among the
three silicon sources persisted throughout the entire 8 wk of study. However, in turkeys, the TES promoted growth much better than NaSil and
NaZA. The finding, in rats, that NaZA and NaSil diets supported superior growth performance, is consistent with previous reports. In studies
involving biological functions of Si, Weigand et al. (10) and Najda et al. (8)
showed higher Si bioavailability from NaZa and NaSil, respectively.
NaZA lowered hemoglobin content in rats. This reduction might be
related more to its A1 content than its Si availability. NaZA contains
about 13% A1. Aluminum has been shown to bind to transferrin and
inhibit Fe transportation by this protein (19). This interference could have
caused a reduction in Fe available for hemoglobin synthesis and, therefore, a drop in hemoglobin values.
The NaZA supplementation of turkeys increased plasma Mg levels,
which supports the report by Birchall (11) that A1 displaces Mg from
tissues and key enzymes, hence the increase in circulating plasma Mg
levels. The NaZA-supplemented turkeys had also significantly enlarged
hearts and livers compared to NaSil supplementation but not TES. This
could again be due to the antagonistic effects of the NaZA-inherent A1
to Fe and Cu, creating anemic and hypocupric conditions. Hypocupria
has been associated with cardiac hypertrophy (20,21). It should be
pointed out that although the hepatic and cardiac tissue Cu concentrations were generally lower for NaZA diets than NaSil and TES, this
reduction was not significantly different. In turkeys, NaZA supplementation significantly reduced liver Mn levels compared to NaSil but not
TES. Carlisle (4) reported that Si depressed Mn bioavailability. In case
of NaZA the antagonistic interaction between Si and Mn could also
have been augmented by the antagonistic effects of the NaZA-inherent
A1 on Mn, a bivalent cation. Birchall (11) reported that A1 displaced
bivalent cations like Ca and Mg. A number of turkeys in this treatment
group seemed to show early symptoms of perosis, a Mn-deficiency disease in birds (22).
Where rats and turkeys responded to the Si sources differently
could be due to the differences in the anatomical structure of the gastrointestinal tracts of the two species. This definitely affects the location
and extent of hydrolysis or solubility of the various compounds and,
therefore, Si absorption. It could also be due to the fact that NaSil and
NaZA are inorganic compounds, whereas TES is an organic compound and, therefore, probably handled differently by the two
species. It should also be pointed that slight variations in total Si
intake by the species might account for some of the observed differences too.

Biological Trace Element Research

Vol. 67, 1999

186

Kayongo-Male and Jia

CONCLUSION
Rats and turkeys fed NaZA diets performed better than those on NaSil
or TES diets. Although NaZA was most effective in promoting higher
growth rates, it also had negative side effects on hemoglobin content, heart
and liver weights, and hepatic Mn concentrations. These side-effects were
deemed to be due to the 13% inherent A1 content of NaZA. Aluminum may
have interfered with the metabolism of other trace elements, such as Fe, Cu,
and Mn. Such negative side effects could have disastrous effects on animal
health when supplemented with NaZA over long periods of time.

ACKNOWLEDGMENTS
We are grateful for the help of Dr. R. J. Emerick and Dr. I. S. Palmer,
the technical support of Renata Wnuk and Nancy Anderson, and the
skillful preparation of the manuscript by Vickie Molengraaf.
This work was funded by the USDA Contract #9404222.

REFERENCES
1. R. K. Iler, The Chemistry of Silica, Wiley, New York, (1979).
2. C. D. Seaborn and F. H. Nielsen, Nutr. Today 28, 13-18 (1993).
3. E. M. Carlisle, Silicon in Trace Elements in Man and Animal Nutrition, Vol. 2, 5th ed.,
W. Mertz, ed., Academic, New York, pp. 373-390 (1986).
4. E. M. Carlisle, Ciiba Found. Symp. 121, 123-139 (1986).
5. E. M. Carlisle, Science 178, 619-621 (1972).
6. R. J. Emerick, Nutr. Rep. Int. 34, 907-913 (1986).
7. R. J. Emerick, J. Nutr. 117, 1924-1928 (1987).
8. J. Najda, J. Gminski, M. Drozdz, and A. Danch, Biol. Trace Element Res. 37, 107-114 (1993).
9. K. Schwarz and D. B. Milne, Nature 239, 333-334 (1972).
10. K. E. Wiegand, Poultry Sci., 70, (Suppl.1), 131 (1991).
11. J. D. Birchall, Silicon and the bioavailability of aluminum--nutritional aspects, in
Food, Nutrition and Chemical Toxicity, D. V. Parke, C. Ioannides, and R. Walker, eds.,
Smith-Gordon, Great Britain, pp. 215-226 (1993).
12. T. P. A. Kruch and D. R. Mclachlan, Mechanisms of aluminum neurotoxity relevance
to human diseases, in Metal Ions in Biological Systems, Vol. 24, H. Sigel and A. Sigel
eds., Marcel Dekker, New York (1988).
13. R. J. Emerick, and H. Kayongo-Male, J. Nutr. Biochem. 1, 35-40 (1990).
14. National Research Council, Nutrient Requirements of Poultry, 9th ed., National Academy
Press, Washington, DC (1994).
15. S. R. Stewart, R. J. Emerick, and H. Kayongo-Male, J. Anim. Sci. 71, 946-954 (1993).
16. B. L. Oser, Hawk's Physiological Chemistry, 14th ed., McGraw-Hill, New York (1965).
17. SAS, SAS Procedures Guide (Release 6.03 Ed.), SAS Institute Inc., Cary, NC (1988).
18. R. G. D. Steel and J. H. Torrie, Principles and Procedures of Statistics: A Biomedical
Approach, 2nd ed., McGraw-Hill, New York (1980).
19. A. C. Alfrey, Aluminum, in Trace Elements in Human and Animal Nutrition, Vol. 2, 5th
ed., W. Mertz, ed., Academic, New York, pp. 399-413 (1986).
20. L. M. Klevy, Ann. NY Acad. Sci. 355, 140-151 (1980).
21. L. M. Klevy, Clin. Geriatr. Med. 3, 361-372 (1987).
22. L. S. Hurley, and C. L. Keen, Manganese, in Trace Elements in Human and Animal
Nutrition, Vol. 1, 5th ed., W. Mertz, ed., Academic, New York, Vol. 1, pp. 185-224 (1987).
Biological Trace Element Research

Vol. 67, 1999