Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 64, 1935
This text is also available in PDF
(1 MB) Opens in new window
– 35 –

The Estimation of Aluminium in Pastures, with Special Reference to Soil Contamination.

[Received by Editor, 28th June, 1933; issued separately, May, 1934.]


In view of the fact that aluminium is the most abundant and widely distributed metallic element of the earth's crust, it is not surprising that it should prove a frequent, if not a universal, constituent of plant and animal tissues. Recent investigations, however, would seem to show that many of the earlier determinations of aluminium in biological materials are too high, much of the aluminium reported having come either from contamination by soil dust or from the traces of aluminium invariably present in the reagents employed.

While it is by no means certain that aluminium is essential either for plant growth or animal nutrition, the estimation of this element is of some importance in connection with the problem of soil contamination—a factor which frequently interferes with the determination of the true content in the pasture of certain trace constituents such as iron, manganese, and iodine. These elements usually occur to a much greater extent in soils than in pastures so that even traces of contaminating material may add very considerably to the amounts reported in the analysis of the mineral content. As many of these trace elements play an important part in animal nutrition, it is most desirable to determine the amounts of these elements actually elaborated by the plant tissues where they are probably present in a more digestible form than in the soil.

Although the greatest care may be exercised in obtaining clean pasture samples, experience has shown that the contamination factor can seldom be disregarded in the interpretation of the ash analysis. Indeed, in certain districts of New Zealand, as, for example, Morton Mains, Southland, the soil contamination of even carefully picked samples is so great as to mask any abnormalities in the true mineral content of the trace elements of the pasture.

For some years past Aston (1928)2 has used the estimation of alumina as a significant aid in determining the degree of contamination by soil or fine atmospheric dust which the pasture sample has undergone. The principle involved depends on the fact that almost invariably the higher plants (to which grasses and clovers belong) absorb only traces of aluminium Jost (1907)7, and hence, if more than traces of this element be found in the pasture sample, this must be attributed to soil or dust contamination.

Before alumina can be used as a quantitative index of contamination it is thus first necessary to determine the limits of the aluminium content of clean grasses and clovers. The excess

– 36 –

aluminium may then be attributed to the adhering material, which might be assumed to have the same mineral content as the finer particles of the soil on which the pasture grows. If the amount and composition of the contaminating material be thus ascertainable, the mineral content actually present in the plant tissues could be found on subtracting this from the mineral content of the contaminated pasture. It is quite possible, however, that the pasture contamination, often consisting largely of atmospheric dust, differs considerably in chemical composition from the soil beneath, and hence wherever practicable the best solution of the problem appears to be the removal of the contaminating material by careful cleaning prior to analysis.

The present investigation was undertaken to obtain further information on some of the problems outlined above, and in particular to establish a satisfactory means of discriminating between the natural pasture minerals and those contributed by the adhering soil. Accordingly, a study was made of the limits of the alumina content of clean pastures and of the relationship of the alumina content to the soil contamination. To facilitate the investigation the gravimetric methods ordinarily employed for the estimation of aluminium were replaced by a convenient colorimetric procedure described below.

Aluminium in Plants.

During the past fifty years numerous investigators have made quantitative estimations of aluminium in plants. The wide variations in the results obtained, however, would seem to suggest that many of the samples analysed were contaminated.

According to the recent work of Bertrand and Levy (1931)3 aluminium was found in all the phanerogams examined in amounts varying from one-tenth of a milligram to several hundred milligrams per kilogram of dry material. Winter and Bird (1929)21 from the analysis of a number of plants and plant materials obtained results of a similar order. Since aluminium was still found to be present in those samples in which the outer skin had been removed, these investigators came to the conclusion that this element is probably a natural constituent of plant and animal tissues.

Generally speaking, the results of earlier investigators indicate that while aluminium occurs throughout all parts of the same plant, it is concentrated chiefly in the roots and is found only in minute quantities in the leaves.5, 15 This, however, is not confirmed by the later work of Bertrand and Levy (1931)3, who found that although edible roots appear to contain far less aluminium than wild roots. the greatest proportion of this element is contained in the leaves and appears to be related to the chlorophyll.4 The inner etiolated leaves of the cabbage, for example, were found to contain only 8 milligrams whereas the outside leaves contained 232 milligrams of aluminium per kilogram of dry material.

The great majority of plants contain only traces of aluminium in unknown combination. In some samples of Lycopodium, however, as much as one-third of the ash is alumina which can be

– 37 –

extracted from the plant in combination with an organic acid.17 Smith19 reported abnormal quantities of alumina in the ash of Orites excelsa R. Br. (Proteaceae), where the aluminium appears to be necessary for growth, any excess being deposited in the cavities and natural fissures of the wood as basic aluminium succinate. In general, whereas xerophytes absorb only traces of aluminium, the hygrophytes are noted for their relatively high aluminium content, especially in the root.20

The Physiological Action of Aluminium on Plants.

According to Stoklasa18 aluminium in very dilute concentrations exerts a favourable influence on seed germination while larger concentrations are toxic. Aluminium in certain suitable concentrations will apparently reduce the toxic effect of excess manganese. With regard to plant growth, the same observer has shown that aluminium is less toxic than iron, and in suitably low concentrations has the effect of reducing the toxicity of excess of either iron or manganese.

In contrast to the detoxicating and stimulating action of very low concentrations of aluminium salts, in recent years a considerable amount of discussion has taken place with regard to the toxicity of soluble aluminium salts present in appreciable amounts in certain acid soils. Line (1926)10 contended that “the toxic aluminium theory” was untenable—the relatively poor growth of plants in soils or nutrient solutions to which aluminium salts have been added (as compared with the control plants) being entirely due to the progressive hydrolysis of the aluminium salts producing increased acidity and a depletion of the phosphate supply owing to the precipitation of aluminium phosphate. Recent work, however, has shown this contention to be groundless. Spencer6, for example, found that the toxicity of aluminium salts to Rhododendron ponticum L. seedlings in sand cultures actually decreased with increasing acidity of the solution. At pH 3.0 a very noticeable stimulating effect occurred with 3 parts per million of aluminium.

In the experiments of McLean and Gilbert12 the plants were placed alternately between complete nutrient solutions containing phosphate and complete nutrient solutions without phosphate, but containing aluminium. The pH of these solutions was kept at 4.0—4.5. Thus the effects of phosphate starvation and increasing acidity were obviated. The results of these studies showed that lettuce, beet, and barley were very sensitive to aluminium, whereas maize, turnips, and redtop were fairly resistant to aluminium poisoning.

The symptoms of aluminium poisoning of plants are apparently seen first in the dwarfing and injury to roots accompanied by a decreased permeability of the plants to dyes or nutrient solutions.

Aluminium in Animals.

Owing to the minute amounts of aluminium usually present in animal tissues and the relative insensitivity of the methods used for its estimation, the results of earlier analysts are not dependable.

– 38 –

Myers and Morrison (1928)13, using a sensitive colorimetric method, found traces of aluminium in the tissues of the dog. Of the tissues analysed—heart, kidney, spleen, and liver—the latter was found to contain the largest amount (15 mgms. Al per 100 gms.). Human autopsy tissue14 was also examined, the figures obtained being of the same order as those found for the dog. Lewis9, using the spectroscopic method, was unable to detect aluminium in human blood, except in certain cases after the subjects had been fed on an aluminium rich diet, when minute traces of this element were indicated. Sheep's blood, on the contrary, was found to contain from 1 to 1½ parts per million of aluminium.

From the numerous metabolism experiments conducted on various animals including man, it would appear that the amount of aluminium absorbed from an aluminium rich diet is very small, and little if any of this element is stored up in the tissues. Aluminium in small amounts is evidently not toxic; very large amounts may, however, produce mild catharsis.


I. The Estimation of Aluminium in Pastures.—Experience in this laboratory has led the author to the conclusion that the gravimetric methods ordinarily employed for the estimation of aluminium, while doubtless accurate for appreciable quantities of this element, are not suitable for pasture analysis where the aliquot solution taken frequently contains less than several milligrams of aluminium. Rather than use an inconveniently large aliquot it seemed preferable to examine the possibilities of the colorimetric methods for the determination of aluminium.

A study of the literature showed that two colorimetric reagents for the estimation of aluminium were available—sodium alizarin sulphonate and aurin tricarboxylic acid. The second reagent has found favour with many analysts, and for some years past the conditions and characteristics of the aluminium lake have been examined. Quite recently Lampitt and Sylvester8 have proposed the use of aurin tricarboxylic acid in conjunction with the Lovibond tintometer for the determination of aluminium in foodstuffs. Their method is claimed to be accurate and, with certain slight modifications embodied in the procedure described below, was found quite suitable for the estimation of aluminium in pastures.

Analytical Procedure.

a. Preparation of the Solution.—For the approximate determination of aluminium the following preliminary preparation is sufficient. Measure a suitable aliquot* of the pasture solution into a 100 c.c. beaker, and take just to dryness on the water bath. Add 2 c.c. 5N hydrochloric acid and 5 c.c. of water. Warm for about one minute and shake to dissolve any precipitate. It will be seen from Table I that this procedure may give a low figure owing to interference by

[Footnote] * Enough to give from 1.8 to 6.0 red units.

– 39 –

other substances. For strictly accurate work the following modification was found satisfactory. Add one drop of methyl orange indicator to the aliquot and dilute to 15 c.c. Exactly neutralise this solution with 5% ammonia, boiling the solution to remove any slight excess of ammonia. Allow the solution to stand for at least an hour and then filter through a Whatman No. 42 paper, washing the precipitate with cold water. Dissolve the precipitate back into the precipitation beaker with hot 25% hydrochloric acid. Take solution just to dryness on the water bath. Add 2 c.c. 5N hydrochloric acid and 5 c.c. water. Warm to dissolve the precipitate.

b. Development of the Colour.—Add 30 c.c. of the colorimetric reagent, mix thoroughly, and place beaker in water bath for 5 minutes. Cool solution in running water for at least 5 minutes. Add 3 c.c. of ammonium hydroxide—carbonate reagent to a 50 c.c. standard flask. Wash the solution into the standard flask and make up to the mark with distilled water, mixing the contents thoroughly.

c. Measurement of the Colour.—Add 30 c.c. of the solution from the standard flask into the Lovibond tintometer tube. Match the colour and read off the red units on the Lovibond scale exactly 5 minutes after neutralisation of the solution with the ammonium hydroxide—carbonate reagent. The amount of aluminium is then ascertained directly from a graph constructed from measurements on known amounts of the standard aluminium solution.


1. 5N hydrochloric acid.

2. 5% ammonia solution.

3. Colorimetric reagent. This must be prepared immediately before each series of determinations. Mix one part by volume of 5N ammonium acetate with 4 parts of 50% glycerine and finally add 1. part of a .002% solution of aurin tricarboxylic acid exactly neutralised with ammonia.

4. Standard aluminium solution. Dissolve 1.757 gm. of pure potash alum Al2(SO4), 3K2 SO4, 24H2 O in distilled water containing 25 c.c. of 5N hydrochloric acid. Make solution up to 1 litre. Dilute 5 times to give a standard solution of which 1 c.c. = 0.02 mgm. of aluminium.

5. Ammonium hydroxide—carbonate solution. Mix equal volumes of 1ON ammonium hydroxide solution and 2N ammonium carbonate. Keep the mixture in a well-stoppered bottle.*

II. Accuracy of the Method.—The accuracy of the method was checked by a comparison with the colorimetric and gravimetric determinations on the same series of pastures solutions and by the recovery of aluminium added in known amounts to a synthetic solution prepared from pure analytical reagents.

[Footnote] * It is recommended that the reagents should be occasionally checked with the standard aluminium solution to ensure that no deterioration has taken place.

– 40 –

Table I.—A Comparison of the Gravimetric and Colorimetric Methods of Estimation of Aluminium in Pastures.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Lab. No. Species. % Alumina.
Gravimetric. Colorimetric.
Determined directly in presence of other mineral constituents. Aluminium first precipitated by ammonia in the presence of phosphate.
7102–3-4 Danthonia (com- 0.046 0.026 0.028
Duplicate posite sample) 0.032 0.025 0.032
7187–8-9 Danthonia (com- 0.055 0.048 0.052
Duplicate posite sample) 0.047 0.046 0.054
7561 General pasture 0.090 0.092
7304 General pasture 0.140 0.090 0.095
Duplicate General pasture 0.152 0.100 0.086
7369 General pasture 0.020 0.024
7373 General pasture 0.016 0.019
4068 Cocksfoot 0.042 0.037 0.040
Duplicate Cocksfoot 0.045 0.038
2102 Red Clover 0.050 0.059
Duplicate Red Clover 0.052 0.064

The gravimetric determinations shown in Table I. involved the precipitation of iron and aluminium with ammonia after the phosphate had been removed. The iron was then determined colorimetrically with thiocyanate and the alumina obtained by difference.

It will be observed that the presence of other substances is inclined to give a low figure for the colorimetric estimation. When, however, the estimation is made on the precipitate obtained by neutralising the solution with ammonia in the presence of phosphates there is fair agreement between the gravimetric and colorimetric methods, the agreement of the duplicates on the whole being closer when the latter method is used.

Table II.—The recovery of aluminium from a synthetic pasture solution containing:—CaO 1.0 per cent., P2 O3 0.75 per cent., K2O 2.62 per cent., Na2O 0.20 per cent., SO3 0.82 per cent., MgO 0.40 per cent., Fe 0.007 per cent., and Mn 0 022 per cent.

Aluminium added Aluminium found
% Al2O3. % Al2O3.
0.095 0.097
0.076 0.080
0.048 0.049
0.038 0.041

The above table shows the method to be accurate to within 5 per cent.

[Footnote] † Weight of precipitate insufficient for accurate determination.

– 41 –

III. Preparation of the Solution for the Determination of Aluminium in Pastures.—The two most common methods of preparing plant solutions for the analysis of their mineral content involve either ashing followed by extraction of the ash with hydrochloric acid or wet digestion with some oxidising agent such as nitric and sulphuric acids. As a result of past experience in the analysis of pasture samples in this laboratory it was known that the first method did not completely extract all the iron and aluminium, a certain proportion of these elements being always retained in the siliceous residue. It was not known, however, whether the wet digestion method would effect a more complete extraction of these elements from the pasture, and to test this point a series of determinations involving both methods were made. The results of this investigation are presented in Tables III and IV.

Table III.—Iron and aluminium occluded in the crude silica of pastures (Ashing Method).

(All results expressed as percentages of the moisture-free grass.)
In solution. Occluded in SiO2 Total.
Lab. No Sample SiO2 Al2O3 Fe Al2O3 Fe Al2O3 Fe
7430 General pasture 4.49 0.057 0.016 0.019 0.0030 0.076 0.019
7427 General pasture 2.90 0.079 0.022 0.026 0.0070 0.105 0.029
8125 Meadow fescue 4 73 0.056 0.006 0.021 0.0035 0.077 0.0095
8102 Browntop 9.15 0.074 0.008 0.021 0.0030 0.095 0.011
4069 Cocksfoot 2.31 0.045 0.008 0.017 0.0033 0.062 0.0135

Table IV.—Iron and aluminium occluded in the crude silica of pastures (Wet Digestion Method).

(All results expressed as percentages of the moisture-free grass.)
In solution. Occluded in SiO2 Total.
Lab. No. Sample SiO3 Al2O3 Fe Al2O3 Fe Al2O3 Fe
4062 Cocksfoot 2.3 0.016 0.005 0.030 0.0035 0.046 0.0085
4069 Cocksfoot 2.4 0.022 0.0075 0.018 0.005 0.040 0.0125
7427 General pasture 3.7 0.042 0.015 0.032 0.007 0.074 0.022
7430 General pasture 4.05 0.036 0.014 0.020 0.004 0.056 0.018
8102 Browntop 8.95 0.025 0.008 0.021 0.004 0.046 0.012
8114 Lotus major 0.23 0.019 0.008 0.015 0.006 0.034 0.014
8125 Meadow fescue 4.6 0.021 0.006 0.021 0.004 0.042 0.010
– 42 –

The results shown in Table III were obtained from pasture samples which had been picked over to remove stalks and obvious contaminating material before grinding in a special bronze mill. The preparation of the pastures for wet digestion involved a preliminary brushing to remove any adhering soil, consequently the alumina figures are somewhat lower than for the ashing method.

It will be seen from the above tables that percentage of siliceous residue and occluded iron and alumina is nearly the same for both the ashing and the wet digestion methods. Approximately 35 to 65 per cent. of the total alumina and 25 to 40 per cent. of the total iron estimated on the cleaned pasture are held up as part of the insoluble residue. It would be interesting to investigate whether the insoluble iron and aluminium occurs as such in the pasture, or whether in the process of wet digestion or of ashing an insoluble silicate is formed. In this connection it is noteworthy that Sjoman16 has shown that when quartz and iron oxide are heated together, the Si atoms in the SiO2 lattice are partly replaced by Fe atoms.

Since the method of wet digestion appeared to offer no advantages over the ashing method, it was decided to adhere to the standard procedure practised in this laboratory. The air-dried pasture was ashed in a porcelain dish at a temperature not exceeding dull red heat, and the residue was taken to dryness twice with hydrochloric acid to render the silica insoluble. Next the mineral constituents were taken up with hydrochloric acid and allowed to stand for 24 hours after which the crude silica was filtered off. The latter was then ignited and again extracted with hydrochloric acid for 24 hours and finally filtered, the filtrate being combined with the first filtrate. In order that the estimations might be comparable with those previously obtained in this laboratory the iron and alumina in the insoluble residue were not estimated.

IV. The Distribution of Aluminium and Iron in Grasses and Legumes.—The distribution of aluminium and iron in grasses and legumes was studied primarily to compare the concentration of these elements in those parts of the plant which are protected from dust contamination, such as the stem of the toetoe (Arundo conspicua) after the outer sheath has been stripped, with the concentration in those parts where contamination is unavoidable, as, for example, the roots and leaves. The data obtained are collected in Table V. All samples were carefully washed with distilled water to remove any adhering soil and, as an extra precaution, the roots in addition to being washed with distilled water were carefully scraped to remove the outer surface.

– 43 –

Table V.—The distribution of aluminium and iron in grasses and legumes.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

(All results expressed as percentages of the moisture-free material.)
Sample Al2O3 Fe Remarks
Roots 0.026 0.010 Local sample, grown on hillside in dry locality.
Stem 0.0025 0.0017
Leaves 0.017 0.0081
Pampas-grass (Gynerium argentium)
Roots 0.052 0.027 Local sample, grown in dry locality.
Stem 0.0031 0.0020
Plume 0.010 0.0043
Leaves 0.009 0.0052
Toetoe (Arundo conspicua)
Stem 0.010 0.0062 Local sample, grown on sandhills.
Plume 0.019 0.0065
Leaves 0.032 0.0125
Root 0.114 0.090 Greenhouse sample.
Leaves 0.010 0.018 Scanty growth.
Red clover—
Root 0.104 0.022 Local sample.
Leaves and stem 0.052 0.018 Rank growth.
White clover—
Stem 0.033 0.011 Local sample, grown on sandy soil.
Leaves 0.035 0.022
Flower 0.085 0.043

It will be observed from Table V that while aluminium occurs in appreciable amounts throughout all parts of the grasses, the highest proportion is found in the roots, the stems being relatively deficient in aluminium as compared with the plumes and leaves. From the limited number of observations made it would seem that clovers differ from grasses in their higher average aluminium content in the above-ground portion.

In both grasses and clovers the distribution of iron appears to follow approximately that of the aluminium.

V. The Effect of Cleaning on the Aluminium and Iron Content of Pastures.—In this investigation the samples representing slightly, medium, and heavily contaminated pastures were taken respectively from the Rotorua, Taranaki, and Morton Mains districts.

After picking over the air-dried pasture to remove woody stalks and obvious contaminating material according to the usual practice of this laboratory, a suitable aliquot (about 40 grams) was set aside and thoroughly brushed to remove as completely as possible any loosely adhering soil and dust. A weighed amount of the brushed

– 44 –

pasture was then placed in a 10-inch porcelain basin and washed with distilled water. After about five minutes' agitation the water generally became discoloured owing to the suspension of clay and soil particles. The whole of the sample was then removed and placed in a clean basin of distilled water, the process being repeated as before until such time as the water showed no appreciable amount of suspended matter. In the case of the Morton Mains pastures, a certain amount of clay adhered very firmly to the grass-blades, necessitating a preliminary soaking for one hour with distilled water.

Table VI.—The effect of cleaning on the alumina and iron content of pastures.

(All results expressed as percentages of the moisture-free pasture.)
Standard Procedure. Pasture brushed. Pasture washed with water.
No. Lab. Sample Al2O3 Fe Al2O3 Fe Al2O3 Fe Locality.
2102 Red clover 0.051 0.009 0.025 0.008 0.017 0.0055
4069 Cocksfoot 0.045 0.008 0.019 0.006 0.006 0.0055
8120 Poa pratensis 0.022 0.010 0.016 0.009 0.007 0.006
8122 Meadow foxtail 0.022 0.009 0.024 0.0085 0.022 0.0085 Kaharoa, Rotorua.
8127 Ratstail 0.029 0.007 0.028 0.007 0.013 0.0045
8137 Chewings fescue 0.049 0.012 0.010 0.011 0.0095 0.011
7427 General pasture 0.079 0.022 0.038 0.013 0.023 0.010 Te Popo, Taranaki,
7430 " 0.057 0.016 0.029 0.010 0.018 0.009
7493 General pasture 0.40 0.083 0.11 0.066 0.039 0.017
7494 " 0.65 0.22 0.18 0.12 0.15 0.062 Morton Mains, Southland.
7495 " 0.16 0.032 0.030 0.018 0.022 0.015
7497 " 0.35 0.21 0.20 0.12 0.051 0.023

It will be seen from the results presented in Table VI that washing is more effective than brushing in the removal of contaminating material from the pasture. It was thought, however, that the former method might not be valid owing to the possible extraction of the pasture minerals.

Accordingly, the solubilities of the chief inorganic constituents were tested as follows. 20 grams of air-dried pasture were shaken up with 500 c.c. of distilled water for 20 hours, 1 c.c. of toluene being added to inhibit bacterial action. The solution was then quickly filtered once through a Whatman No. 41 filter paper and finally through a No. 42 paper. The results of this investigation are collected together in Table VII.

– 45 –

Table VII.—The solubility of the chief inorganic constituents of pastures in water.

(All results expressed as percentages of the moisture-free pasture.)
Total Mineral Content. Water Soluble Mineral Content. pH Indicactor Mthd. Water sble, acids pr 100 gms. Airdried pst [ unclear: ] e
Lab. No. Species CaO P2O5 K2O Fe Al2O3 Mn MgO CaO P2O5 K2O Fe Al2O3 Mn MgO
4068 Cocksfoot 0.41 0.72 3.16 0.06 0.044 0.046 0.34 0.27 0.55 2.64 0.0007 none 0.0052 0.17 5.5 = 19.9 c.c. N. NaOH.
7427 General pasture 1.41 0.40 1.84 0.022 0.079 0.014 0.49 0.32 0.32 1.50 0.0009 trace 0.0026 0.22 5.0 = 18.7 c.c. N. NaOH.
7430 " 0.78 0.42 1.72 0.016 0.057 0.009 0.41 0.10 0.36 1.70 0.0006 " 0.0034 0.18 5.5 = 16.2 c.c. N.NaOH.
8114 Lotus major 0.90 0.61 3.18 0.012 0.074 0.011 0.70 0.28 0.42 2.63 trace none 0.0018 0.37 4.5 = 26.7 c.c. N. NaOH.
8120 Poa pratensis 0.56 0.65 2.00 0.010 0.074 0.018 0.45 0.26 0.50 1.80 0.0004 " 0.0039 0.30 5.5 = 18.5 c.c. N. NaOH.
8137 Chewings fescue 0.58 0.71 2.43 0.012 0.049 0.046 0.23 0.24 0.54 2.08 0.0008 " 0.0086 0.17 6.0 = 18.2 c.c. N. NaOH.
– 46 –

Table VIII.—The solubility of the chief inorganic constituents of pastures in water.

(Each result expressed as a percentage of the coriesponding total inorganic constituent.)
Potash Magnesia Manganese Aluminium Phosphate Lime Iron
(K2O) (MgO) (Mn) (A1) (P2O5) (CaO) (Fe)
82–99% 44–74% 11–38% trace % 69–86% 12–66% trace-11%

The practice of steeping pastures in water to remove earthy material evidently leads to the extraction of some of the major constituents as is shown from Table VIII. So far as iron is concerned, however, this element is only slightly soluble, so that washing of the pasture should not extract any appreciable amount, especially if the process occupies only a few minutes. Moreover, it is unlikely that any iron is dissolved out from the pasture in the ferrous state and subsequently oxidised and precipitated in the solution as ferric hydroxide, since solubility experiments conducted entirely in an atmosphere of nitrogen gave the same results for the solubility of iron as those conducted in air. Thus while the method of washing the pasture renders the sample unsuitable for a general analysis of the mineral content, it affords a satisfactory means of estimating the true pasture iron in a contaminated pasture sample.

A noteworthy feature arising from Table VII is the remarkable difference in the solubility of the same mineral constituent in different pastures, particularly in the case of lime. This would seem to suggest that a pasture with a high lime content might not necessarily contain as much digestible lime as a pasture with much less total lime.

VI. The Availability of the Pasture Iron in Pepsin-Hydrochloric Acid Solutions.—As it had been suggested by Askew and Rigg1 that bush sickness might be caused by a soil iron deficiency rather than by a deficiency of iron in the pasture, it seemed desirable to obtain some data regarding the digestibility of iron in contaminated and uncontaminated pastures. In the absence of facilities for the carrying out of metabolism experiments a series of contaminated and uncontaminated pastures were digested in pepsin-hydrochloric acid solution. The experiments were conducted as follows. A weighed amount of air-dried pasture, approximately 5 grams, was placed in a beaker with 400 c.c. of 0.35 per cent. hydrochloric acid solution containing 0.2 per cent. pepsin and warmed for six hours at a

– 47 –

temperature of 37° C. Throughout the experiment the solution was stirred to secure a more complete reaction. At the end of the period the solution was filtered and evaporated to dryness. After igniting the extract, the ash was analysed in the usual way. The preparation of the pastures for the pepsin-hydrochloric acid consisted of a preliminary picking over to remove stalks and obvious contaminating material according to the usual practice, except that the air-dried grass was not ground.

Table IX.—The digestibility of iron in contaminated and uncontaminated pasture samples.

(All results expressed as percentages of the moisture-free grass.)
Standard procedure. Cleaned pasture. Soluble in Pepsin hydrochloric acid.
Lab. No. Species Fe Fe Fe Locality
8114 Lotus major 0.008 0.007 Kaharoa,
8120 Poa pratensis 0.010 0.006 0.0035 Rotorua.
8137 Chewings fescue 0.012 0.011 0.004
7427 General pasture 0.022 0.010 0.012 Te Popo,
7430 " " 0.016 0.009 0.0075 Taranaki.
7493 General pasture 0.083 0.017 0.012
7494 " " 0.22 0.062 0.055 Morton
7495 " " 0.032 0.015 0.006 Mains,
7497 " " 0.21 0.023 0.035 Southland.

Generally speaking, the results shown in Table IX indicate that the percentage of iron soluble in pepsin-hydrochloric acid is high, ranging above 35 per cent. of the true pasture iron. The most striking feature of the table, however, is the fact that even excess iron contamination adds but little to the amount of digestible iron which thus appears to be closely related to the natural iron of the pasture.

In the absence of actual metabolism experiments, it may be tentatively concluded that the digestibility of the pasture iron varies considerably according to the species of the grass, and that probably but little of the contaminating iron is available to the animal.

General Discussion.

The results obtained on carefully cleaned pastures support the conclusion that most of the aluminium ordinarily found in pastures is due to soil contamination. This is a matter of some importance in the assessing of the food value of the iron, as pepsin-hydrochloric acid digestion experiments on contaminated and uncontaminated pastures suggest that whereas the natural pasture iron is readily soluble, the iron contributed by soil contamination is by no means

– 48 –

easily digestible, and hence the interpretation of the mineral content analysis requires that some allowance be made for the contamination factor.

The alumina content of clean pastures generally falls below 0.025 per cent. of the moisture-free pasture, and the excess alumina above this limit might therefore be attributed to soil contamination. It must be remembered, however, that acidity favours the absorption of iron and aluminium by the plant, and that while the limiting factor of 0.025 per cent. alumina might be given to Te Popo and Kaharoa pastures where the acidity of the soil usually ranges round about pH 5.7 and pH 5.5 respectively, such a factor might not apply to soils whose acidity falls outside the minimum solubility range for alumina defined by Magistad11 as extending from pH 4.7 to pH 8.5.

Attention should perhaps be drawn to the excessive iron and alumina content of some of the Morton Mains pastures; one carefully cleaned sample, for example, contained approximately ten times the normal amount of both iron and aluminium. Moreover, it is probable that the iron and aluminium reported was not due to contamination, since further cleaning of the pasture failed to effect any appreciable reduction in the amounts of these elements. Further experiments are being made to account for this abnormality. In the meantime, it is tentatively suggested that the relatively high acidity (pH 5.1) of the Morton Mains soil from which this last pasture sample was taken might account for the abnormal absorption of iron and aluminium.

Summary and Conclusions.

(1) A modification of the Lampitt and Sylvester method suitable for the estimation of aluminium in pastures is described.

(2) The modified method is found to be accurate to within about 5 per cent., the agreement of duplicate determinations being closer by this method than by the gravimetric method ordinarily used.

(3) The destruction of the organic matter in pastures either by ashing or by nitric and sulphuric digestion leaves from 35 to 65 per cent. of the total pasture alumina and from 25 to 40 per cent. of the total pasture iron in the insoluble residue.

(4) The distribution of aluminium in the bamboo, pampas-grass, toetoe, ryegrass, red clover, and white clover has been studied, the amounts of alumina found varying from 0.0025 per cent. in bamboo stems to 0.114 per cent. in ryegrass roots.

– 49 –

(5) The alumina of the grasses studied is shown to be concentrated chiefly in the roots, the plumes and leaves containing more alumina than the stems.

(6) The figures obtained on carefully cleaned pastures show that, generally speaking, the alumina content is below 0.025 per cent., supporting the idea that most of the alumina ordinarily found is due to soil contamination.

(7) Alumina appears to be the best index of soil contamination in pasture samples, especially contamination from fine atmospheric dust which may not be easily detectable by other methods.

(8) In the estimation of iron the problem of soil contamination is best overcome by brushing and careful washing of the pasture prior to ashing.

(9) Experiments show that this process does not involve any appreciable loss of pasture iron. The washing of pastures renders them unsuitable for a general estimation of the mineral content as the major ingredients of the pasture, i.e., K2O, P2O5, CaO, MgO, are so combined in the plant as to be largely extractable by water.

(10) Digestion of contaminated and uncontaminated pastures in pepsin-hydrochloric acid shows that the amount of soluble iron, which usually exceeds 35 per cent. of the true pasture iron, is closely related to the amount of true pasture iron and is practically independent of the amount of contaminating iron.


The author wishes to express his thanks to Mr B. C. Aston, chief chemist, Department of Agriculture, for permission to carry out the work, and to Mr R. E. R. Grimmett for his interest and advice in the interpretation of the results.


1. Askew, H.O., and Rigg. T.: Bulletin No. 32, N.Z. Dept. of Sci. and Ind. Res. (1932), 20

2. Aston, B. C.; N.Z. J. Agric. (1928), 36, 24.

3. Bertrand, G., and Levy. G.; Bull. Soc. Chem. (1931), 49–50, 1417–1425.

4. Bertrand, G., and Mrs G. Levy; Amer. Chem. Abs. (1931), 25, 3686. Compt. rend. Acad. Agr. France (1931), 17, 235–8.

5. Berthelot and Andre, G.; Compt. rend. (1895), 120, No. 6, 288–290.

6. Brenchley, W. E.: J. Agric Sci. (1932), 22, 726.

7. Jost, L.; Lectures on Plant Physiology. Translated by R. J. H. Gibson, Oxford, at the Clarendon Press (1907), p. 86.

8. Lampitt, L. H., and Sylvester, N. D.; Analyst (1932), 57, No. 676, 418–426.

9. Lewis, S. J.; Biochem. J. (1931), 25, 2162–2167.

– 50 –

10. Line, J.; J. Agric. Sci. (1926), 16, 335.

11. Magistad, O. C.; J. Agric. Sci. (1925), 20, 181–225.

12. McLean, F. T., and Gilbert, B. E.; Soil Sci. (1927), 24, 163–174.

13. Myers, V. C., and Morison, D. B.; J. Biol. Chem. (1928), 78, 615–624.

14. Myers, V. C., and Mull, J. W.; J. Biol. Chem. (1928), 78, 625–626.

15. Ricciardi, L.; Gaz. Chim. Ital. (1889), 19, 150–159.

10. Sjoman, P.; B. C. A. (1931), A. 1010; Teku. Samf. Handl. (1930), No. 7, 24 pp., Chem. Lentr. (1931), 1, 1737.

17. Smith, E. E.; Aluminium Compounds in Food. P. B. Hoeber Inc., New York (1928), p. 13.

18. Smith, E. E.; Ibid., 176–179.

19. Smith, M. G.; J. and Proc. Roy. Soc., New South Wales (1903), 37, 107–120.

20. Stoklasa, J.; International Rev. Sci. Prac. Agric. (1925), 3, 655–662.

21. Winter, O. B., and Bird, A. O.; J. Amer. Chem. Soc. (1929), 51, 2964–2968.