- Email: [email protected]
Biological Availability of Manganese Sources and Effects of High Dietary Manganese on Tissue Mineral Composition of Broiler-Type Chicks1,2
Biological Availability of Manganese Sources and Effects of High DietaryManganese on Tissue Mineral Composition of Broiler-Type Chicks1'2 J. R. BLACK,3 C. B. AMMERMAN,4 P. R. HENRY, and R. D. MILES 5 Department of Animal Science University of Florida, Gainesville, Florida 32611 (Received for publication January 19, 1984) ABSTRACT An experiment was conducted with male broiler-type chicks to study tissue uptake of Mn as a measure of biological availability of Mn sources. A basal corn-soybean meal diet (116 ppm Mn) was supplemented with 0, 1000, 2000, or 4000 ppm Mn as either reagent grade sulfate, carbonate, or monoxide and fed ad libitum for 26 days. No toxic effects were noted as expressed by feed intake, weight gain, feed conversion, hematocrit, hemoglobin, or mortality. Analysis of Mn in tissues revealed a highly linear relationship between liver or bone Mn concentration and dietary Mn for all three sources. Manganese concentration in all tissues increased (P<.01) as dietary Mn increased. Liver and bone Mn accumulation appeared to be excellent indicators of relative biological availability. On the basis of tissue uptake and solubility tests, M n S 0 4 - H 2 0 was the most available, followed by MnO and MnC0 3 , respectively. There were no effects of Mn source or level on concentration of Ca, P, Mg, Cu, Zn, or Fe in tissues studied. (Key words: manganese, tissue minerals, biological availability) 1984 Poultry Science 63:1999-2006
INTRODUCTION
Relative biological availability of Mn sources is a practical and economic concern when feeds inherently low in Mn such as corn and sorghum grain, are a major portion of poultry diets (Underwood, 1981). Criteria used to determine biological availability include incidence of perosis (Gallup and Norris, 1939) and leg bone abnormality (Watson et al., 1970; 1971). Sulfate and chloride forms of Mn have been found to be more available than the carbonate or dioxide forms. It has been suggested (Miller, 1983) that biological availability of minerals may be determined more precisely by measuring some
1 Florida Agricultural Experiment Station Journal Series No. 5300. 2 The authors wish to acknowledge International Minerals and Chemical Corp., Mundelein, IL, Moorman Manufacturing Co., Quincy, IL, Southeastern Minerals, Inc., Bainbridge, GA, and Occidental Chemical Co., Houston, TX for funds in support of this research, S. M. Free, Smith-Kline Animal Health Products, Philadelphia, PA for assistance with the statistical analysis and Agricultural Operations, Minerals Division, International Minerals & Chemical Corp., Bartow, FL for X-ray diffraction analysis of Mn sources. 3 Central Soya Co., 1200 N. 2nd St., Decatur, IN 46733. 4 To whom reprint requests should be addressed. 'Department of Poultry Science.
aspect of functional effectiveness such as growth rate, remission or prevention of deficiency signs, and biochemical changes when sources are added to deficient diets. There are certain disadvantages to the use of deficiency signs, such as subjectively quantifying effects (e.g., leg bone deformity score), the increased cost of purified diets, and possibility of errors due to contamination. Using tissue Mn accumulation at high but nontoxic levels as a measure of biological availability has the advantages that natural diets may be fed that allow animals to reach full genetic growth potential, any contamination would be relatively insignificant, and modern atomic absorption techniques are sufficiently sensitive to quantify tissue Mn concentrations accurately. The Select Committee on GRAS (Generally Recognized As Safe) Substances of the Life Sciences Research Office (1979) reported in an extensive review that while MnO supplies 90% of the Mn in animal feeds, there are insufficient data upon which to evaluate it as a GRAS feed ingredient. No. acute oral toxicity studies or long-term feeding studies of MnO were available. The following study was conducted to investigate tissue uptake at high dietary Mn levels as an objective method for determining relative biological availability of Mn from different sources and the effect of high Mn levels on tissue mineral concentration in chicks.
1999
2000
BLACK ET AL. TABLE 1. Composition of basal diet
Item
(%) Dietary ingredients1 Corn, ground yellow Soybean meal, dehulled Poultry oil Dicalcium phosphate (18.5% P; 22% Ca) Limestone, ground Microingredients2 Salt, iodized DL-Methionine Filler, washed sand 3 Total Chemical composition" Crude protein, % Metabolizable energy, kcal/kg Ca, % P, % Mg, % Mn, ppm Fe, ppm Zn, ppm Cu, ppm 1
55.65 37 2.50 1.70 1 .50 .40 .25 1 100 23 3016 .91 .76 .18 116 580 58 12.5
As-fed basis
2
Ingredients supplied per kilogram of diet: vitamin A palmitate, 6600 IU; vitamin D 3 , 2200 ICU; menadione dimethylpyrimidinol bisulfite, 2.2 mg; riboflavin, 4.4 mg; pantothenic acid, 13.2 mg; niacin, 39.6 mg; choline chloride, 499.4 mg; vitamin B 1 2 , 22 ixg; ethoxyquin, .0125%; manganese, 60 mg; iron, 50 mg; copper, 6 mg; zinc, 36 mg. 'Manganese supplement added at the expense of equivalent weights of washed sand. "Dry matter basis (88.1% dry matter in diet).
chicks National Research Council (NRC, 1977). Treatments included reagent grade sources of M n S 0 4 - H 2 0 , MnC0 3 , and MnO added to the basal diet at 0, 1000, 2000, or 4000 ppm Mn at the expense of washed sand. At the termination of the experiment, final chick weights and feed consumption were determined and weight gain calculated. Four chicks from each pen were randomly selected for tissue analysis. Blood samples were taken by anterior cardiac puncture, and plasma was saved for mineral analysis. Hemoglobin was determined by a cyanmethemoglobin method (Unopette test 5857/5858; Becton-Dickinson, Rutherford, NJ) and hematocrit by a microhematocrit method (Cohen, 1967). Chicks were killed by cervical dislocation and liver, pectoralis major muscle, heart, and right tibia were excised for mineral analysis. Calcium, Mg, Cu, Fe, Zn, and Mn in tissues, diets, plasma, and Mn sources were determined by flame atomic absorption spectrophotometry (AAS) on a Perkin-Elmer Model 5000 (Anonymous, 1982), except for plasma Mn, which was determined by flameless AAS using a Perkin-Elmer 503 with HGA-2100 graphite furnace (Anonymous, 1974). Phosphorus in Mn sources, diets, tissues, and plasma was determined by a modified colorimetric method (Harris and Popat, 1954). Relative solubilities (Watson et al., 1970) and magnetic susceptibility (Watson et al., 1971) of the Mn sources were determined, and X-ray diffraction patterns were interpreted. All data were analyzed by analysis of variance (Steel and Torrie, 1980), and differences were separated by Duncan's (1955) multiple range test.
MATERIALS AND METHODS
One hundred and twenty, day-old male Cobb color-sexed chicks were used in a completely randomized design with a 3 X 4 factorial arrangement of treatments. Chicks were randomly assigned to pens in a thermostatically-controlled, electrically-heated Petersime battery with raised wire floors. Two replicates of six chicks were assigned to each treatment and fed a basal corn-soybean meal diet to 4 days posthatching. Chicks were maintained on a 24-hr, constant-light schedule and allowed ad lib. access to feed and tap water. On Day 4 posthatching, chicks were weighed and fed the experimental diets for 26 days. The basal diet (Table 1) was a corn-soybean meal diet (116 ppm Mn) formulated to meet or exceed the nutrient requirements of growing
RESULTS AND DISCUSSION Analysis of mineral sources (Table 2 and 3) indicated Mn concentrations of 32.6, 47.1, and 76.7 for the sulfate, carbonate, and oxide sources, respectively. The reagent grade Mn sources were relatively free of mineral contaminants. The sulfate was 100% soluble in water and the carbonate and oxide were virtually insoluble. The carbonate was only 30.3% as soluble in neutral ammonium citrate as the sulfate and oxide sources and all forms were completely soluble in dilute (.4%) HC1. The oxide was slightly insoluble in 2% citric acid. Solubility did not appear to be related to particle size. Only the oxide form had significant magnetic susceptibility and it did not
MANGANESE AVAILABILITY
2001
TABLE 2. Chemical characteristics of manganese sources Chemical constituents, dry m a t t e r basis Sources
Mn
P
Ca
Mg
32.6 47.1 76.7
Fe
22.7 15.7 8.3
118 186 4.8
Particle size, % Cu
+ 30 1
- 3 0 + 100
-100
1.2
52.2 4.8 38.1
46.6 95.2 61.9
,
(%) Sulfate Carbonate Oxide
Zn
(ppm) 2
44.7 47.0 21.2
51.8 53.0 50.2
1
Retained by a No. 30 sieve (US Bureau of Standards).
2
None detected.
appear to be related to Fe contamination. A high degree of purity was indicated by Xray diffraction. There were no differences (P>.05) in feed intake, average daily gain, or feed conversion (56.3 ± .37 g, 30.4 ± .17 g, and 1.85 ± .02, respectively) due to treatment, which indicated that there was no gross toxicosis due to the elevated dietary Mn. Mortality was not related to treatment, because only one death occurred in one replicate of the 2000 ppm oxide treatment. Hemoglobin and hematocrit averaged 7.75 ± .11 g/dl and 28.3 ± .22% red blood cells, respectively, and were not affected by supplemental Mn. Plasma Mn increased (P<.01) from 3.74 £ig/dl as dietary Mn increased for all sources, although that for the 1000 ppm level was not different from controls (Table 4). The 4000 ppm dietary Mn as sulfate produced higher (P<.01) Mn levels in the plasma (13.6
2
748
/ig/dl) than the equivalent level from carbonate (9.1 Mg/dl), oxide (12.0 jug/dl) was intermediate. Regression analysis (Table 5) indicated an inconsistent, but linear, increase in plasma Mn as dietary levels increased. The sulfate produced the most linear response in plasma as shown by the high correlation coefficient (r = .92). Plasma Ca, P, and Mg averaged 11.5 ± .13, 7.55 ± .05, and 1.99 ± .03 mg/dl and Zn and Cu 223 ± 3.4 and 18.7 ± .42 jug/dl and were not affected by dietary Mn. Liver Mn increased (P<.01) as dietary Mn increased from all sources (Table 4). Regression analysis (Table 5) indicated a highly linear uptake of Mn by liver as dietary Mn increased with all sources tested. Liver Mn concentrations at the 1000 ppm level from sulfate (20.9 ppm) and oxide (17.8 ppm) were higher (P<.01) than the unsupplemented control (11.6 ppm) and the 4000 ppm level from these sources produced higher liver Mn (31.0 and 27.4 ppm)
TABLE 3. Physical characteristics of manganese sources Relative solubility 1
Sources
Physical appearance
Magnetic I n t e r p r e t a t i o n suscepti- of X-ray bility patterns
Water
Neutral ammonium citrate
.4% HC1
2% Citric acid
100
100
100
100
2.3
MnS04-H20 (reagent grade) M n C 0 3 (reagent grade)
Sulfate
Light pink, crystalline p o w d e r
Carbonate
Light tan, fine p o w d e r
.8
30
100
100
1.8
Oxide
Dark green, fine granules
.4
100
100
92
84.6
1
Relative solubility at 37 C for 1 hr, constant stirring.
MnO (reagent grade)
3.74 6.34 8.41 13.57 6.45 7.57 9.10 6.25 10.87 12.04
+ 22a ± 29ab ± 99bc + 72e ± 69ab + 42bc ± 71 bed + 12ab ± 1 03cde + 22de
(Mg/dO
Plasma2
11.6 20.9 24.2 31.0 16.1 18.9 20.3 17.8 22.4 27.4
± ± + ± ± ± ± ± ± ±
.61a 59bcd 1 13de 2 31f 53ab 70bc 88bcd 71bc 89cd 2 ll^f
Liver, dry basis2
2.15 3.39 3.73 3.89 3.06 2.93 2.61 2.93 2.40 3.14
2
Mean ± SE of four chicks per treatment for plasma and eight chicks per treatment for other tissues.
Basal diet contained 116 ppm Mn; chicks fed 26 days.
(ppm)±.22a ± 28ab ± 36bc ± 23c ± 28abc ± 43abc + 29abc ± 31abc + 31ab ± 34abc
Heart, dry basis2
' ' ' ' ' Means within columns with different superscripts differ significantly (P<.01).
Sulfate Sulfate Sulfate Carbonate Carbonate Carbonate Oxide Oxide Oxide
Mn source
>y>zMeans within columns with different superscripts differ significantly (P<.05).
1
x
0 1000 2000 4000 1000 2000 4000 1000 2000 4000
(ppm)
Added Mn
.79 1.10 1.71 2.25 1.02 1.72 2.31 1.21 1.40 1.52
± ± ± ± ± ± + + ± +
.04a 06ab 20bc 20c 07ab 26bc 35c 16ab 13ab 17abc
Muscle, dry basis2
TABLE 4. Effect of source and level of dietary manganese on manganese concentration of tissues and
MANGANESE AVAILABILITY T A B L E 5. Linear regression analysis of tissue manganese with respect to dietary manganese Mn source
Correlation coefficient
Reg ression e q u a t i o n Plasma 1 Y = 3.75 + . 0 0 2 4 4 X Y = 4.52 + .00126X Y = 4.50 + .00212X
Sulfate Carbonate Oxide
.92 .78 .87
Liver 2 Y = 13.93 + . 0 0 4 5 6 X Y = 13.11 + . 0 0 2 0 6 X Y = 13.09+ .00383X
Sulfate Carbonate Oxide
.86 .80 .85
Heart 2 Y = Y = Y =
Sulfate Carbonate Oxide
2.61 + . 0 0 0 3 8 6 X 2.57 + . 0 0 0 0 7 1 X 2.32 + . 0 0 0 1 9 6 X
.57 .11 .33
Muscle 2 Sulfate Carbonate Oxide
Y = Y = Y =
.81 + . 0 0 0 3 7 6 X .75 + .OO04O3X .33 + . 0 0 0 1 6 8 X
.81 .71 .57
Bone2 Y = 15.41 + . 0 4 8 6 X Y = 17.88 + . 0 1 6 1 X Y = 21.47 + .0279X
Sulfate Carbonate Oxide
.95 .88 .90
1 Where Y equals plasma Mn, Mg/dl, and X equals dietary Mn, p p m . Each regression e q u a t i o n represents 16 chicks. 2 W h e r e Y equals tissue Mn, p p m , and X equals dietary Mn, p p m . Each regression e q u a t i o n represents 32 chicks.
than the same level from the carbonate source (20.3 ppm). Liver Ca, P, Mg, Zn, Fe, and Cu were not affected by dietary treatment (Table 6).
T A B L E 6. Effect Tissue
Ca
Liver Heart Muscle
111 190 77
of source and level of dietary
2003
Dietary Mn increased heart Mn concentration (Table 4), but only the 2000 and 4000 ppm treatment from sulfate (3.73 and 3.89 ppm) produced tissue Mn concentrations higher (P<.01) than the control group (2.15 ppm). The heart showed inconsistent uptake of Mn. Skeletal muscle (Table 4) contained lower concentrations of Mn than cardiac muscle and increased (less than 2 ppm) as dietary Mn increased. Only the 2000 and 4000 ppm levels from sulfate and carbonate produced muscle Mn concentrations that were higher than controls (P<.01). A more linear response was found for skeletal muscle than for cardiac muscle, which is evident from linear regression analysis of the two tissues (Table 5). There were no differences in other heart or muscle minerals analyzed (Table 6). Bone ash was decreased by some treatments (P<.05), but no consistent pattern was evident (Table 4). Manganese in bone ash increased (P<.01) in a highly linear manner (Table 4) as indicated by regression analysis (Table 5) from all Mn sources. The sulfate produced greater Mn concentrations than the oxide or carbonate at both the 2000 and 4000 ppm level of supplementation (P<.01). The oxide produced higher bone Mn than the carbonate form at the 4000 ppm level (P<.01). Bone Mn was increased from 12.3 ppm in controls to 207.4, 78.6, and 129.8 ppm for the sulfate, carbonate, and oxide sources when supplemented at 4000 ppm. This amounts to greater than a 16-fold increase for the sulfate source. Bone Ca, Mg, and P were not affected by dietary Mn (Table 6). Results of the present study indicate that tissue uptake of Mn, when sources are fed to chicks at high but nontoxic concentrations,
manganese
on mineral
Mg
composition
Zn
of tissues in
chicks1
Fe
Cu
256 ± 8.7 222 ± 7.4 14 ± .41
14.9 ±.24 19.6 ±.39 3.2 ±.14
NA
NA
( p p m , dry m a t t e r basis) ±4.0 ± 3.2 + 2.4
10822 6783 8657
± 103 ± 81 ± 262
648 ± 4.3 808 ± 8.3 978 ±13.8
85 ± .82 97 ±1.20 19 ± .24
(%, ash weight basis) Bone
37.1
.17
17.4
±
.08
7 2 4 ± 83
-
NA2
' Values represent m e a n ± SE of 8 0 chicks fed 2 6 days. P > . 1 0 for t r e a t m e n t c o m b i n a t i o n s . 2
N o t analyzed.
2004
BLACK ET AL.
could be used to differentiate biological availability among sources. Tissue Mn concentration was increased (P<.01) in all tissues studied when the element was supplemented up to 4000 ppm. When dietary Mn from M n S O ^ H 2 0 was increased 35-fold bone Mn increased 16-fold, and liver increased 2.7-fold. Manganese concentrations in bone (Watson et al., 1970; Southern and Baker, 1983b) and in bile (Southern and Baker, 1983a) had been shown previously to increase linearly with increasing dietary levels. The linear uptake of Mn by liver and bone makes these tissues useful measures of biological availability that are more precisely measured than leg bone deformity scores (Watson et al., 1970; 1971). Differences in Mn bioavailability are more readily apparent in tissue when Mn is fed at high levels than with assays using purified diets supplemented at levels below the requirement. This is indicated by higher correlation coefficients for bone (r = .95, .88, and .90 for sulfate, carbonate, and oxide, respectively) in the present study compared to r = .67 when smaller increments (0, 10, 20, 30 ppm Mn) were added to purified diets as M n S 0 4 - H 2 0 (Watson et al, 1970). Skeletal and cardiac muscles were variable in Mn uptake and contained relatively low Mn concentrations. Thus, muscle uptake would not be a satisfactory measure of biological availability. Solubility of Mn in neutral ammonium citrate was highly correlated with biological
availability in chicks (Watson et al, 1971). By this criterion, the sulfate and oxide sources should be more available than the carbonate, as confirmed by tissue uptake. Based on bone and liver uptake and solubility, MnS04*H 2 0 had the highest availability followed by MnO, with MnC03 the least available of the three reagent grade sources tested. Estimates of relative value of the Mn sources may be determined in several ways. Multiple linear regression of bone ash Mn concentration as related to dietary levels yields the following equation: Y = 22.20 + .046362XJ + .014649X2 + .027617X3 (r= .995). [1] where Y equals bone Mn, ppm ash basis, Xj equals ppm Mn as sulfate, X 2 equals ppm Mn as carbonate, and X 3 equals ppm Mn as oxide. Similarly, multiple linear regression of liver Mn concentration with respect to dietary Mn yields: Y = 14.56 + .004352Xi + .001570X 2 + .003339X3 (r = .968) [2] where Y equals liver Mn, ppm dry basis, and other variables as in [1]. Estimates of relative biological availability (Table 7) may be obtained by a ratio of the slopes in equations [1] and [2] , which represents the rate of change in tissue Mn with
TABLE 7. Relative biological availability of manganese sources based on multiple linear regression, linear regression, and tissue manganese concentration
Source
Multiple linear regression slope
Relative value
Linear regression slope
(%) Sulfate Carbonate Oxide
.0464 .0146 .0276
100.0 31.6 59.6
Relative value
Average tissue Mn increase
(%) Bone .0486 .0161 .0279
100.0 33.1 5 7.4
Relative value
(%) 117.6 1 44.9 77.2
100.0 38.2 65.6
Liver Sulfate Carbonate Oxide
.00435 .00157 .00334
1
Ash weight basis (ppm).
2
Dry matter basis (ppm).
100.0 36.1 76.7
.00456 .00206 .00383
100.0 45.2 84.0
13.76 2 6.81 10.92
100.0 49.5 79.4
2005
MANGANESE AVAILABILITY respect to change in dietary Mn. Setting MnSC>4 • H 2 0 at 100% gives 31.6% for MnC0 3 and 59.6% for MnO using bone Mn concentrations and 36.1 and 76.7% using liver Mn concentrations. Alternatively, slopes from individual regression lines (Table 5) may be compared in a similar manner. This gives relative biological availability estimates of 3 3.1 and 57.4% for MnC0 3 and MnO using bone or 45.2 and 84.0% for liver (Table 7). If average increase in tissue Mn concentration over basal is used for comparison, slightly higher estimates are obtained, 38.2% for MnC0 3 and 65.6% for MnO in bone or 49.5 and 79.4% in liver. All estimates yield the same rank order for the different sources. An index may be calculated using the mean of the three individual estimates in Table 7 to approximate relative value of the sources. The estimated index values for bone are 34.3% for MnC0 3 and 60.9% for MnO. The estimated index values for liver are slightly higher, 43.6% for MnC0 3 and 80.0% for MnO. In this experiment bone produced a more conservative estimate index. Based on leg bone deformity score of chicks fed 10 ppm supplemental Mn in purified diets, Watson et al. (1971) found a reagent grade MnS04*H 2 0 more available than rhodocrosite (a natural carbonate ore), but not different from reagent grade M n C 0 3 . The sulfate was also more available than four oxide sources tested. In a second experiment with the same sources and feeding levels, leg deformity score could not differentiate statistically among any of the sources. Southern and Baker (1983a) used multiple regression of bile Mn concentration on Mn intake at 0, 3000, and 4000 ppm to yield an equation considering sulfate, carbonate, chloride, and dioxide sources. Using slope ratio and setting M n S 0 4 ' H 2 0 as 100% gave bioavailability estimates of 77% for M n C 0 3 , 102% for MnCl2 ' 4 H 2 0 , and 29% for M n 0 2 . The present study indicated that reagent grade MnO (with no measurable M n 0 2 ) was more available than MnC0 3 . This is consistent with earlier studies that found availability of oxides related to relative proportions of MnO to M n 0 2 (Watson etal, 1971). Although there were no signs of gross Mn toxicosis in the present study, several trends suggested adverse effects at the highest level of supplementation. These included numerically reduced hemoglobin and hematocrit as all
sources of dietary Mn were increased and numerically decreased dietary intake at the highest level of sulfate supplementation. The present study using modern poultry diets agrees with other studies, indicating that Mn toxicosis from highly available sources such as MnS0 4 * H 2 0 will occur at between 4000 and 5000 ppm Mn when growth depression is the criterion (Vohra and Kratzer, 1968; Southern and Baker, 1983a). Higher concentrations could be tolerated with less available sources. As the monoxide source appeared to be somewhat less available and, therefore, less toxic than Mn sulfate, it should be considered for the list of GRAS substances. Liver and bone Mn uptake were sensitive and objective methods of determining relative biological availability of Mn sources when fed at high but nontoxic levels. Data indicated that M n S 0 4 * H 2 0 was more available than MnO, and MnC0 3 was the least available of the three sources tested. Further studies will be conducted with natural-type diets supplemented with Mn sources at concentrations above and below NRC (1977) Mn requirement levels to determine if the linear response in tissue uptake observed at high dietary Mn additions is consistent at lower levels of supplementation. REFERENCES Anonymous, 1974. Analytical Methods for Atomic Absorption Spectroscopy Using the HGA Graphite Furnace. Perkin-Elmer Corp., Norwalk, CT. Anonymous, 1982. Analytical Methods for Atomic Absorption Spectrophotometry. Perkin-Elmer Corp., Norwalk, CT. Cohen, R. R., 1967. Anticoagulation, centrifugation time, and sample replicate number in the microhematocrit method for avian blood. Poultry Sci. 46:214-218. Duncan, D. B., 1955. Multiple range and multiple F tests. Biometrics 1 1 : 1 - 4 2 . Gallup, W. D., and L. C. Norris, 1939. The amount of manganese required to prevent perosis in the chick. Poultry Sci. 18:76-82. Harris, W. D., and P. Popat, 1954. Determination of the phosphorus content of lipids. J. Am. Oil Chem. Soc. 31:124-130. Miller, W. J., 1983. Biological availability of trace mineral elements. Proc. 1983 Nutr. Inst. Min., Natl. Feed Ingred. Assoc. West Des Moines, IA. National Research Council, 1977. Nutrient Requirements of Domestic Animals. 1. Nutrient Requirements of Poultry. 7th ed. Natl. Acad. Sci., Washington, DC. Select Committee on GRAS Substances, 1979. Evaluation of the health aspects of manganous salts as food ingredients (SCOGS-67). Life Sci. Res. Office, Fed. Am. Soc. Exp. Biol., Bethesda, MD. Southern, L. L., and, D. H. Baker, 1983a. Excess
2006
BLACK ET AL.
manganese ingestion in the chick. Poultry. Sci. 62:642-646. Southern, L. L., and D. H. Baker, 1983b. Eimeria acervulina infection in chicks fed deficient or excess levels of manganese. J. Nutr. 113:172— 177. Steel, R.G.D., and J. H. Torrie, 1980. Principals and Procedures of Statistics: A Biometrical Approach. 2nd ed. McGraw-Hill Book Co., New York, NY. Underwood, E. J., 1981. The Mineral Nutrition of Livestock. 2nd ed. Commonw. Agric. Bur.
Slough, England. Vohra, P., and F. H. Kratzer, 1968. Zinc, copper and manganese toxicities in turkey poults and their alleviation by EDTA. Poultry Sci. 47:699-704. Watson, L. T., C. B. Ammerman, S. M. Miller, and R. H. Harms, 1970. Biological assay of inorganic manganese for chicks. Poultry Sci. 49:1548— 1554. Watson, L. T., C. B. Ammerman, S. M. Miller, and R. H. Harms, 1971. Biological availability to chicks of manganese from different inorganic sources. Poultry Sci. 50:1693-1700.
INTRODUCTION
Relative biological availability of Mn sources is a practical and economic concern when feeds inherently low in Mn such as corn and sorghum grain, are a major portion of poultry diets (Underwood, 1981). Criteria used to determine biological availability include incidence of perosis (Gallup and Norris, 1939) and leg bone abnormality (Watson et al., 1970; 1971). Sulfate and chloride forms of Mn have been found to be more available than the carbonate or dioxide forms. It has been suggested (Miller, 1983) that biological availability of minerals may be determined more precisely by measuring some
1 Florida Agricultural Experiment Station Journal Series No. 5300. 2 The authors wish to acknowledge International Minerals and Chemical Corp., Mundelein, IL, Moorman Manufacturing Co., Quincy, IL, Southeastern Minerals, Inc., Bainbridge, GA, and Occidental Chemical Co., Houston, TX for funds in support of this research, S. M. Free, Smith-Kline Animal Health Products, Philadelphia, PA for assistance with the statistical analysis and Agricultural Operations, Minerals Division, International Minerals & Chemical Corp., Bartow, FL for X-ray diffraction analysis of Mn sources. 3 Central Soya Co., 1200 N. 2nd St., Decatur, IN 46733. 4 To whom reprint requests should be addressed. 'Department of Poultry Science.
aspect of functional effectiveness such as growth rate, remission or prevention of deficiency signs, and biochemical changes when sources are added to deficient diets. There are certain disadvantages to the use of deficiency signs, such as subjectively quantifying effects (e.g., leg bone deformity score), the increased cost of purified diets, and possibility of errors due to contamination. Using tissue Mn accumulation at high but nontoxic levels as a measure of biological availability has the advantages that natural diets may be fed that allow animals to reach full genetic growth potential, any contamination would be relatively insignificant, and modern atomic absorption techniques are sufficiently sensitive to quantify tissue Mn concentrations accurately. The Select Committee on GRAS (Generally Recognized As Safe) Substances of the Life Sciences Research Office (1979) reported in an extensive review that while MnO supplies 90% of the Mn in animal feeds, there are insufficient data upon which to evaluate it as a GRAS feed ingredient. No. acute oral toxicity studies or long-term feeding studies of MnO were available. The following study was conducted to investigate tissue uptake at high dietary Mn levels as an objective method for determining relative biological availability of Mn from different sources and the effect of high Mn levels on tissue mineral concentration in chicks.
1999
2000
BLACK ET AL. TABLE 1. Composition of basal diet
Item
(%) Dietary ingredients1 Corn, ground yellow Soybean meal, dehulled Poultry oil Dicalcium phosphate (18.5% P; 22% Ca) Limestone, ground Microingredients2 Salt, iodized DL-Methionine Filler, washed sand 3 Total Chemical composition" Crude protein, % Metabolizable energy, kcal/kg Ca, % P, % Mg, % Mn, ppm Fe, ppm Zn, ppm Cu, ppm 1
55.65 37 2.50 1.70 1 .50 .40 .25 1 100 23 3016 .91 .76 .18 116 580 58 12.5
As-fed basis
2
Ingredients supplied per kilogram of diet: vitamin A palmitate, 6600 IU; vitamin D 3 , 2200 ICU; menadione dimethylpyrimidinol bisulfite, 2.2 mg; riboflavin, 4.4 mg; pantothenic acid, 13.2 mg; niacin, 39.6 mg; choline chloride, 499.4 mg; vitamin B 1 2 , 22 ixg; ethoxyquin, .0125%; manganese, 60 mg; iron, 50 mg; copper, 6 mg; zinc, 36 mg. 'Manganese supplement added at the expense of equivalent weights of washed sand. "Dry matter basis (88.1% dry matter in diet).
chicks National Research Council (NRC, 1977). Treatments included reagent grade sources of M n S 0 4 - H 2 0 , MnC0 3 , and MnO added to the basal diet at 0, 1000, 2000, or 4000 ppm Mn at the expense of washed sand. At the termination of the experiment, final chick weights and feed consumption were determined and weight gain calculated. Four chicks from each pen were randomly selected for tissue analysis. Blood samples were taken by anterior cardiac puncture, and plasma was saved for mineral analysis. Hemoglobin was determined by a cyanmethemoglobin method (Unopette test 5857/5858; Becton-Dickinson, Rutherford, NJ) and hematocrit by a microhematocrit method (Cohen, 1967). Chicks were killed by cervical dislocation and liver, pectoralis major muscle, heart, and right tibia were excised for mineral analysis. Calcium, Mg, Cu, Fe, Zn, and Mn in tissues, diets, plasma, and Mn sources were determined by flame atomic absorption spectrophotometry (AAS) on a Perkin-Elmer Model 5000 (Anonymous, 1982), except for plasma Mn, which was determined by flameless AAS using a Perkin-Elmer 503 with HGA-2100 graphite furnace (Anonymous, 1974). Phosphorus in Mn sources, diets, tissues, and plasma was determined by a modified colorimetric method (Harris and Popat, 1954). Relative solubilities (Watson et al., 1970) and magnetic susceptibility (Watson et al., 1971) of the Mn sources were determined, and X-ray diffraction patterns were interpreted. All data were analyzed by analysis of variance (Steel and Torrie, 1980), and differences were separated by Duncan's (1955) multiple range test.
MATERIALS AND METHODS
One hundred and twenty, day-old male Cobb color-sexed chicks were used in a completely randomized design with a 3 X 4 factorial arrangement of treatments. Chicks were randomly assigned to pens in a thermostatically-controlled, electrically-heated Petersime battery with raised wire floors. Two replicates of six chicks were assigned to each treatment and fed a basal corn-soybean meal diet to 4 days posthatching. Chicks were maintained on a 24-hr, constant-light schedule and allowed ad lib. access to feed and tap water. On Day 4 posthatching, chicks were weighed and fed the experimental diets for 26 days. The basal diet (Table 1) was a corn-soybean meal diet (116 ppm Mn) formulated to meet or exceed the nutrient requirements of growing
RESULTS AND DISCUSSION Analysis of mineral sources (Table 2 and 3) indicated Mn concentrations of 32.6, 47.1, and 76.7 for the sulfate, carbonate, and oxide sources, respectively. The reagent grade Mn sources were relatively free of mineral contaminants. The sulfate was 100% soluble in water and the carbonate and oxide were virtually insoluble. The carbonate was only 30.3% as soluble in neutral ammonium citrate as the sulfate and oxide sources and all forms were completely soluble in dilute (.4%) HC1. The oxide was slightly insoluble in 2% citric acid. Solubility did not appear to be related to particle size. Only the oxide form had significant magnetic susceptibility and it did not
MANGANESE AVAILABILITY
2001
TABLE 2. Chemical characteristics of manganese sources Chemical constituents, dry m a t t e r basis Sources
Mn
P
Ca
Mg
32.6 47.1 76.7
Fe
22.7 15.7 8.3
118 186 4.8
Particle size, % Cu
+ 30 1
- 3 0 + 100
-100
1.2
52.2 4.8 38.1
46.6 95.2 61.9
,
(%) Sulfate Carbonate Oxide
Zn
(ppm) 2
44.7 47.0 21.2
51.8 53.0 50.2
1
Retained by a No. 30 sieve (US Bureau of Standards).
2
None detected.
appear to be related to Fe contamination. A high degree of purity was indicated by Xray diffraction. There were no differences (P>.05) in feed intake, average daily gain, or feed conversion (56.3 ± .37 g, 30.4 ± .17 g, and 1.85 ± .02, respectively) due to treatment, which indicated that there was no gross toxicosis due to the elevated dietary Mn. Mortality was not related to treatment, because only one death occurred in one replicate of the 2000 ppm oxide treatment. Hemoglobin and hematocrit averaged 7.75 ± .11 g/dl and 28.3 ± .22% red blood cells, respectively, and were not affected by supplemental Mn. Plasma Mn increased (P<.01) from 3.74 £ig/dl as dietary Mn increased for all sources, although that for the 1000 ppm level was not different from controls (Table 4). The 4000 ppm dietary Mn as sulfate produced higher (P<.01) Mn levels in the plasma (13.6
2
748
/ig/dl) than the equivalent level from carbonate (9.1 Mg/dl), oxide (12.0 jug/dl) was intermediate. Regression analysis (Table 5) indicated an inconsistent, but linear, increase in plasma Mn as dietary levels increased. The sulfate produced the most linear response in plasma as shown by the high correlation coefficient (r = .92). Plasma Ca, P, and Mg averaged 11.5 ± .13, 7.55 ± .05, and 1.99 ± .03 mg/dl and Zn and Cu 223 ± 3.4 and 18.7 ± .42 jug/dl and were not affected by dietary Mn. Liver Mn increased (P<.01) as dietary Mn increased from all sources (Table 4). Regression analysis (Table 5) indicated a highly linear uptake of Mn by liver as dietary Mn increased with all sources tested. Liver Mn concentrations at the 1000 ppm level from sulfate (20.9 ppm) and oxide (17.8 ppm) were higher (P<.01) than the unsupplemented control (11.6 ppm) and the 4000 ppm level from these sources produced higher liver Mn (31.0 and 27.4 ppm)
TABLE 3. Physical characteristics of manganese sources Relative solubility 1
Sources
Physical appearance
Magnetic I n t e r p r e t a t i o n suscepti- of X-ray bility patterns
Water
Neutral ammonium citrate
.4% HC1
2% Citric acid
100
100
100
100
2.3
MnS04-H20 (reagent grade) M n C 0 3 (reagent grade)
Sulfate
Light pink, crystalline p o w d e r
Carbonate
Light tan, fine p o w d e r
.8
30
100
100
1.8
Oxide
Dark green, fine granules
.4
100
100
92
84.6
1
Relative solubility at 37 C for 1 hr, constant stirring.
MnO (reagent grade)
3.74 6.34 8.41 13.57 6.45 7.57 9.10 6.25 10.87 12.04
+ 22a ± 29ab ± 99bc + 72e ± 69ab + 42bc ± 71 bed + 12ab ± 1 03cde + 22de
(Mg/dO
Plasma2
11.6 20.9 24.2 31.0 16.1 18.9 20.3 17.8 22.4 27.4
± ± + ± ± ± ± ± ± ±
.61a 59bcd 1 13de 2 31f 53ab 70bc 88bcd 71bc 89cd 2 ll^f
Liver, dry basis2
2.15 3.39 3.73 3.89 3.06 2.93 2.61 2.93 2.40 3.14
2
Mean ± SE of four chicks per treatment for plasma and eight chicks per treatment for other tissues.
Basal diet contained 116 ppm Mn; chicks fed 26 days.
(ppm)±.22a ± 28ab ± 36bc ± 23c ± 28abc ± 43abc + 29abc ± 31abc + 31ab ± 34abc
Heart, dry basis2
' ' ' ' ' Means within columns with different superscripts differ significantly (P<.01).
Sulfate Sulfate Sulfate Carbonate Carbonate Carbonate Oxide Oxide Oxide
Mn source
>y>zMeans within columns with different superscripts differ significantly (P<.05).
1
x
0 1000 2000 4000 1000 2000 4000 1000 2000 4000
(ppm)
Added Mn
.79 1.10 1.71 2.25 1.02 1.72 2.31 1.21 1.40 1.52
± ± ± ± ± ± + + ± +
.04a 06ab 20bc 20c 07ab 26bc 35c 16ab 13ab 17abc
Muscle, dry basis2
TABLE 4. Effect of source and level of dietary manganese on manganese concentration of tissues and
MANGANESE AVAILABILITY T A B L E 5. Linear regression analysis of tissue manganese with respect to dietary manganese Mn source
Correlation coefficient
Reg ression e q u a t i o n Plasma 1 Y = 3.75 + . 0 0 2 4 4 X Y = 4.52 + .00126X Y = 4.50 + .00212X
Sulfate Carbonate Oxide
.92 .78 .87
Liver 2 Y = 13.93 + . 0 0 4 5 6 X Y = 13.11 + . 0 0 2 0 6 X Y = 13.09+ .00383X
Sulfate Carbonate Oxide
.86 .80 .85
Heart 2 Y = Y = Y =
Sulfate Carbonate Oxide
2.61 + . 0 0 0 3 8 6 X 2.57 + . 0 0 0 0 7 1 X 2.32 + . 0 0 0 1 9 6 X
.57 .11 .33
Muscle 2 Sulfate Carbonate Oxide
Y = Y = Y =
.81 + . 0 0 0 3 7 6 X .75 + .OO04O3X .33 + . 0 0 0 1 6 8 X
.81 .71 .57
Bone2 Y = 15.41 + . 0 4 8 6 X Y = 17.88 + . 0 1 6 1 X Y = 21.47 + .0279X
Sulfate Carbonate Oxide
.95 .88 .90
1 Where Y equals plasma Mn, Mg/dl, and X equals dietary Mn, p p m . Each regression e q u a t i o n represents 16 chicks. 2 W h e r e Y equals tissue Mn, p p m , and X equals dietary Mn, p p m . Each regression e q u a t i o n represents 32 chicks.
than the same level from the carbonate source (20.3 ppm). Liver Ca, P, Mg, Zn, Fe, and Cu were not affected by dietary treatment (Table 6).
T A B L E 6. Effect Tissue
Ca
Liver Heart Muscle
111 190 77
of source and level of dietary
2003
Dietary Mn increased heart Mn concentration (Table 4), but only the 2000 and 4000 ppm treatment from sulfate (3.73 and 3.89 ppm) produced tissue Mn concentrations higher (P<.01) than the control group (2.15 ppm). The heart showed inconsistent uptake of Mn. Skeletal muscle (Table 4) contained lower concentrations of Mn than cardiac muscle and increased (less than 2 ppm) as dietary Mn increased. Only the 2000 and 4000 ppm levels from sulfate and carbonate produced muscle Mn concentrations that were higher than controls (P<.01). A more linear response was found for skeletal muscle than for cardiac muscle, which is evident from linear regression analysis of the two tissues (Table 5). There were no differences in other heart or muscle minerals analyzed (Table 6). Bone ash was decreased by some treatments (P<.05), but no consistent pattern was evident (Table 4). Manganese in bone ash increased (P<.01) in a highly linear manner (Table 4) as indicated by regression analysis (Table 5) from all Mn sources. The sulfate produced greater Mn concentrations than the oxide or carbonate at both the 2000 and 4000 ppm level of supplementation (P<.01). The oxide produced higher bone Mn than the carbonate form at the 4000 ppm level (P<.01). Bone Mn was increased from 12.3 ppm in controls to 207.4, 78.6, and 129.8 ppm for the sulfate, carbonate, and oxide sources when supplemented at 4000 ppm. This amounts to greater than a 16-fold increase for the sulfate source. Bone Ca, Mg, and P were not affected by dietary Mn (Table 6). Results of the present study indicate that tissue uptake of Mn, when sources are fed to chicks at high but nontoxic concentrations,
manganese
on mineral
Mg
composition
Zn
of tissues in
chicks1
Fe
Cu
256 ± 8.7 222 ± 7.4 14 ± .41
14.9 ±.24 19.6 ±.39 3.2 ±.14
NA
NA
( p p m , dry m a t t e r basis) ±4.0 ± 3.2 + 2.4
10822 6783 8657
± 103 ± 81 ± 262
648 ± 4.3 808 ± 8.3 978 ±13.8
85 ± .82 97 ±1.20 19 ± .24
(%, ash weight basis) Bone
37.1
.17
17.4
±
.08
7 2 4 ± 83
-
NA2
' Values represent m e a n ± SE of 8 0 chicks fed 2 6 days. P > . 1 0 for t r e a t m e n t c o m b i n a t i o n s . 2
N o t analyzed.
2004
BLACK ET AL.
could be used to differentiate biological availability among sources. Tissue Mn concentration was increased (P<.01) in all tissues studied when the element was supplemented up to 4000 ppm. When dietary Mn from M n S O ^ H 2 0 was increased 35-fold bone Mn increased 16-fold, and liver increased 2.7-fold. Manganese concentrations in bone (Watson et al., 1970; Southern and Baker, 1983b) and in bile (Southern and Baker, 1983a) had been shown previously to increase linearly with increasing dietary levels. The linear uptake of Mn by liver and bone makes these tissues useful measures of biological availability that are more precisely measured than leg bone deformity scores (Watson et al., 1970; 1971). Differences in Mn bioavailability are more readily apparent in tissue when Mn is fed at high levels than with assays using purified diets supplemented at levels below the requirement. This is indicated by higher correlation coefficients for bone (r = .95, .88, and .90 for sulfate, carbonate, and oxide, respectively) in the present study compared to r = .67 when smaller increments (0, 10, 20, 30 ppm Mn) were added to purified diets as M n S 0 4 - H 2 0 (Watson et al, 1970). Skeletal and cardiac muscles were variable in Mn uptake and contained relatively low Mn concentrations. Thus, muscle uptake would not be a satisfactory measure of biological availability. Solubility of Mn in neutral ammonium citrate was highly correlated with biological
availability in chicks (Watson et al, 1971). By this criterion, the sulfate and oxide sources should be more available than the carbonate, as confirmed by tissue uptake. Based on bone and liver uptake and solubility, MnS04*H 2 0 had the highest availability followed by MnO, with MnC03 the least available of the three reagent grade sources tested. Estimates of relative value of the Mn sources may be determined in several ways. Multiple linear regression of bone ash Mn concentration as related to dietary levels yields the following equation: Y = 22.20 + .046362XJ + .014649X2 + .027617X3 (r= .995). [1] where Y equals bone Mn, ppm ash basis, Xj equals ppm Mn as sulfate, X 2 equals ppm Mn as carbonate, and X 3 equals ppm Mn as oxide. Similarly, multiple linear regression of liver Mn concentration with respect to dietary Mn yields: Y = 14.56 + .004352Xi + .001570X 2 + .003339X3 (r = .968) [2] where Y equals liver Mn, ppm dry basis, and other variables as in [1]. Estimates of relative biological availability (Table 7) may be obtained by a ratio of the slopes in equations [1] and [2] , which represents the rate of change in tissue Mn with
TABLE 7. Relative biological availability of manganese sources based on multiple linear regression, linear regression, and tissue manganese concentration
Source
Multiple linear regression slope
Relative value
Linear regression slope
(%) Sulfate Carbonate Oxide
.0464 .0146 .0276
100.0 31.6 59.6
Relative value
Average tissue Mn increase
(%) Bone .0486 .0161 .0279
100.0 33.1 5 7.4
Relative value
(%) 117.6 1 44.9 77.2
100.0 38.2 65.6
Liver Sulfate Carbonate Oxide
.00435 .00157 .00334
1
Ash weight basis (ppm).
2
Dry matter basis (ppm).
100.0 36.1 76.7
.00456 .00206 .00383
100.0 45.2 84.0
13.76 2 6.81 10.92
100.0 49.5 79.4
2005
MANGANESE AVAILABILITY respect to change in dietary Mn. Setting MnSC>4 • H 2 0 at 100% gives 31.6% for MnC0 3 and 59.6% for MnO using bone Mn concentrations and 36.1 and 76.7% using liver Mn concentrations. Alternatively, slopes from individual regression lines (Table 5) may be compared in a similar manner. This gives relative biological availability estimates of 3 3.1 and 57.4% for MnC0 3 and MnO using bone or 45.2 and 84.0% for liver (Table 7). If average increase in tissue Mn concentration over basal is used for comparison, slightly higher estimates are obtained, 38.2% for MnC0 3 and 65.6% for MnO in bone or 49.5 and 79.4% in liver. All estimates yield the same rank order for the different sources. An index may be calculated using the mean of the three individual estimates in Table 7 to approximate relative value of the sources. The estimated index values for bone are 34.3% for MnC0 3 and 60.9% for MnO. The estimated index values for liver are slightly higher, 43.6% for MnC0 3 and 80.0% for MnO. In this experiment bone produced a more conservative estimate index. Based on leg bone deformity score of chicks fed 10 ppm supplemental Mn in purified diets, Watson et al. (1971) found a reagent grade MnS04*H 2 0 more available than rhodocrosite (a natural carbonate ore), but not different from reagent grade M n C 0 3 . The sulfate was also more available than four oxide sources tested. In a second experiment with the same sources and feeding levels, leg deformity score could not differentiate statistically among any of the sources. Southern and Baker (1983a) used multiple regression of bile Mn concentration on Mn intake at 0, 3000, and 4000 ppm to yield an equation considering sulfate, carbonate, chloride, and dioxide sources. Using slope ratio and setting M n S 0 4 ' H 2 0 as 100% gave bioavailability estimates of 77% for M n C 0 3 , 102% for MnCl2 ' 4 H 2 0 , and 29% for M n 0 2 . The present study indicated that reagent grade MnO (with no measurable M n 0 2 ) was more available than MnC0 3 . This is consistent with earlier studies that found availability of oxides related to relative proportions of MnO to M n 0 2 (Watson etal, 1971). Although there were no signs of gross Mn toxicosis in the present study, several trends suggested adverse effects at the highest level of supplementation. These included numerically reduced hemoglobin and hematocrit as all
sources of dietary Mn were increased and numerically decreased dietary intake at the highest level of sulfate supplementation. The present study using modern poultry diets agrees with other studies, indicating that Mn toxicosis from highly available sources such as MnS0 4 * H 2 0 will occur at between 4000 and 5000 ppm Mn when growth depression is the criterion (Vohra and Kratzer, 1968; Southern and Baker, 1983a). Higher concentrations could be tolerated with less available sources. As the monoxide source appeared to be somewhat less available and, therefore, less toxic than Mn sulfate, it should be considered for the list of GRAS substances. Liver and bone Mn uptake were sensitive and objective methods of determining relative biological availability of Mn sources when fed at high but nontoxic levels. Data indicated that M n S 0 4 * H 2 0 was more available than MnO, and MnC0 3 was the least available of the three sources tested. Further studies will be conducted with natural-type diets supplemented with Mn sources at concentrations above and below NRC (1977) Mn requirement levels to determine if the linear response in tissue uptake observed at high dietary Mn additions is consistent at lower levels of supplementation. REFERENCES Anonymous, 1974. Analytical Methods for Atomic Absorption Spectroscopy Using the HGA Graphite Furnace. Perkin-Elmer Corp., Norwalk, CT. Anonymous, 1982. Analytical Methods for Atomic Absorption Spectrophotometry. Perkin-Elmer Corp., Norwalk, CT. Cohen, R. R., 1967. Anticoagulation, centrifugation time, and sample replicate number in the microhematocrit method for avian blood. Poultry Sci. 46:214-218. Duncan, D. B., 1955. Multiple range and multiple F tests. Biometrics 1 1 : 1 - 4 2 . Gallup, W. D., and L. C. Norris, 1939. The amount of manganese required to prevent perosis in the chick. Poultry Sci. 18:76-82. Harris, W. D., and P. Popat, 1954. Determination of the phosphorus content of lipids. J. Am. Oil Chem. Soc. 31:124-130. Miller, W. J., 1983. Biological availability of trace mineral elements. Proc. 1983 Nutr. Inst. Min., Natl. Feed Ingred. Assoc. West Des Moines, IA. National Research Council, 1977. Nutrient Requirements of Domestic Animals. 1. Nutrient Requirements of Poultry. 7th ed. Natl. Acad. Sci., Washington, DC. Select Committee on GRAS Substances, 1979. Evaluation of the health aspects of manganous salts as food ingredients (SCOGS-67). Life Sci. Res. Office, Fed. Am. Soc. Exp. Biol., Bethesda, MD. Southern, L. L., and, D. H. Baker, 1983a. Excess
2006
BLACK ET AL.
manganese ingestion in the chick. Poultry. Sci. 62:642-646. Southern, L. L., and D. H. Baker, 1983b. Eimeria acervulina infection in chicks fed deficient or excess levels of manganese. J. Nutr. 113:172— 177. Steel, R.G.D., and J. H. Torrie, 1980. Principals and Procedures of Statistics: A Biometrical Approach. 2nd ed. McGraw-Hill Book Co., New York, NY. Underwood, E. J., 1981. The Mineral Nutrition of Livestock. 2nd ed. Commonw. Agric. Bur.
Slough, England. Vohra, P., and F. H. Kratzer, 1968. Zinc, copper and manganese toxicities in turkey poults and their alleviation by EDTA. Poultry Sci. 47:699-704. Watson, L. T., C. B. Ammerman, S. M. Miller, and R. H. Harms, 1970. Biological assay of inorganic manganese for chicks. Poultry Sci. 49:1548— 1554. Watson, L. T., C. B. Ammerman, S. M. Miller, and R. H. Harms, 1971. Biological availability to chicks of manganese from different inorganic sources. Poultry Sci. 50:1693-1700.