Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 64, 1935
This text is also available in PDF
(2 MB) Opens in new window
– 191 –

Some Characteristics of “Limonites” Used in the Cure and Prevention of Bush Sickness.

Received by the Editor, May 10, 1934; issued separately, September, 1934.


The hydrated oxides of iron occurring in nature have been variously and somewhat indefinitely classified both mineralogically and chemically. Posnjak and Merwin (1919) list the following series as having been described as minerals, stating that goethite alone has been well defined:—

  • Turgite 2 Fe2O3 H2O.

  • GoethiteFe 2O3 H2O.

  • Lepidocrocite 2O3 H2O.

  • Hydrogoethite 3 Fe2O3 4 H2O.

  • Limonite 3 Fe2O3 3 H2O.

  • Xanthosiderite 3 Fe2O3 2 H2O.

  • Limnite 3 Fe2O3 3 H2O.

There are also two anhydrous oxides, haematite (Fe2O3) and magnetite (Fe3O4), of which the latter does not concern us here. Clarke (1924) does not recognise hydrogoethite, while Friend (1921) also lists esmeraldaite Fe2O3 4 H2O.

From their optical, physical, and chemical studies of the various natural and synthetic hydrous oxides Posnjak and Merwin conclude that only the monohydrate has a definite chemical existence, occurring in the two crystalline forms of goethite and lepidocrocite, and with various amounts of adsorbed water, in amorphous form as limonite and the other supposed higher hydrates. They also find turgite to be a solid solution of hydrous haematite and the monohydrate. Weiser (1926) concurs in these views and concludes that the yellow crystalline monohydrate prepared by these authors is identical with goethite. He also considers that the yellow hydrated oxide prepared by Tommasi by the oxidation of hydrous ferrous oxide is the monohydrate. This yellow synthetic hydrated oxide does not lose water on prolonged boiling at 100° C. and is only sparingly soluble in concentrated acids. The red-brown hydrous oxide obtained by precipitating a ferric salt with alkali is very soluble in dilute acids, but is dehydrated even by boiling water. A brick-red oxide is produced by prolonged boiling of ferric acetate. This modification is nearly insoluble in concentrated nitric or hydrochloric acids, but “dissolves” (sol formation) in dilute acids, being reprecipitated by the addition of concentrated acid.

Posnjak and Merwin (1919), working with selected type specimens containing generally less than 5 per cent. of impurities showed that in many cases the geological classification did not agree

– 192 –

with that found by analysis, the discrepancies being generally increased by allowing for the water of hydration and impurities such as silica.

With these facts in mind the difficulty of determining the exact mineralogical status of the hydrous iron oxide in the various deposits and preparations of New Zealand “limonites” used for stock feeding purposes, and which generally contain from 10 to 20 per cent. of impurities, can be appreciated. It is also apparent that the mode of formation may greatly influence the solubilities of hydrated iron oxides of approximately the same chemical composition. Naturally occurring hydrated iron oxides may be formed in several ways, as, for instance, by the oxidation and leaching of iron sulphides, precipitation from solution by alkaline waters, and the oxidation of ferrous carbonate by iron bacteria.


Ore from four different sources has been ground and used in experiments on the cure and prevention of bush sickness. That from Ruatangata (Whangarei), much of which is earthy in texture and of light brownish-yellow colour, when crushed and screened with only a preliminary air-drying, has given uniformly good results. *

Ore from Puhipuhi of more compact texture and darker colour appears to have given some good results when treated similarly to the Ruatangata product, but when finely ground in a cement mill during which process it was heated (as shown by the reddish colour and low combined water and from information received) and became mixed with several per cent. of calcium carbonate, it was generally a failure.

Onekaka ore which is compact and dark coloured has not yet been tried in a merely air-dried and ground condition. Some which is known to have been heated was definitely unsuccessful in preventing bush sickness.

Feeding experiments with Okaihau ore have not yet proceeded far enough to indicate its relative efficiency when compared with the Ruatangata material, but some favourable reports have been received. The ground product is a duller brown than any of the others, but the analysis shows it to be high in combined water.

[Footnote] * For a full account of the field evidence and previous literature, see article on “Control of Bush Sickness in Sheep,” by B. C. Aston, N.Z. Jour. Agric., June, 1932.

[Footnote] † Since the above was written, Okaihau finely ground, air dried limonite has been found by Mr C. R. Taylor to be fully efficacious in preventing bush sickness in sheep.

– 193 –

Complete analyses of the Puhipuhi and Onekaka ores have been made by Mr Seelye and are given herewith by courtesy of the Dominion Analyst:—

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

Fusion Analysis of Onekaka and Puhipuhi Ores by Mr Seelye.
Onekaka. Puhipuhi
Silica (SiO2) 17.25 6.33
Alumina (Al2 O3) 3.55 2.54
Ferric oxide (Fe2O3) 65.65 71.25
Magnesia (MgO) 0.25 0.08
Calcium oxide (CaO) 1.35 0.65
Sodium oxide (Na2O) trace trace
Potassium oxide (K2O) 0.53 0.02
Phosphorus pentoxide (P2O5) 0.36 0.40
Titanium dioxide (TiO2) 0.30 0.40
Manganese dioxide (MnO2) 0.16 2.45
Chromium sesquioxide (Cr2O3) 0.05 none
Cupric oxide (CuO) trace not found
Nickelous oxide (NiO) trace trace
Cobaltous oxide (CoO) none trace
Antimony trioxide (Sh2O3) none 0.04
Arsenic trioxide (As2O3) 0.006 0.033
Barium oxide (BaO) trace 0.27
Sulphur (S) 0.14 0.16
Carbon dioxide (Co2) 0.68 0.49
Moisture 1.10 3.20
Loss on ignition 9.15 12.12
100.52 100.43
Oxygen correction 0.18
FeO 100.34

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

Fusion analysis of Ruatangata and Okaihau samples made in this laboratory are as follows:—
A/1115. D/1397.
Ruatangata. Okaihau.
Silica (SiO2) 7.79 2.17
Alumina (Al2O3) 5.24 7.16
Ferric oxide (Fe2O3) 62.30 67.45
Alkalis (K2O, Na2O) trace trace
Calcium oxide (CaO) 0.58 0.56
Magnesia (MgO) 0.05 0.13
Titanium dioxide (TiO2) 0.31 0.62
Manganese dioxide (MnO2) 0.79 0.24
Chromium sesquioxide (Cr2O3) not found not found
Cupric oxide (CuO) 0.08 0.07
Nickelous oxide (NiO) 0.01 0.01
Cobaltous oxide (CoO) trace trace
Antimony trioxide (Sb2O3) not found not found
Arsenic trioxide (As2O3) trace trace
Sulphur (S) not found 0.16
Phosphorus pentoxide (P2O5) 0.82 1.79
Moisture 9.59 5.88
Loss on ignition 13.01 14.07
100.57 100.31

It does not appear that such differences between the ores as are shown in these analyses are adequate to account for the great variations in their curative properties.

– 194 –

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

The Chemical Composition and Characteristics of Some Typical Limonites used in Stock Feeding.
Lab No. %Moisture (dried at 100° C. Lab, No. 24 hours) %Loss on Ignition. Combined water calculated to 100% hydrated iron oxide. %Iron Oxide: % Calcium carbonate (CaCo3) limestone (titration) method. Solubility in sugar hydrochloric acid solution. % Fe2O3 dissolved. Colour. Ridgway's “Colour Standards.” 1912. Remarks.
Fe2O3 FeO.
D/771 2.82 11.85 12.5 72.6 0.05 3.4 9.4 Between Argus and Sudan brown Farmer “H,” Atiamuri
D/772 2,92 12.40 13.4 71.7 trace 2.9 0.69 Sudan brown Farmer “H,” Atiamuri
D/773 1.68 10.63 10.7 73.8 trace 4.11 1.14 Argus brown Farmer “P,” Putaruru Puhipuhi limonites. Reported inefficient in field trials. (Ground in cement mill.)
D/774 3.87 13.50 14.3 72.6 0.23 3.2 0.84 Antique brown Farner “N,” Tokoroa
D/838 2.58 14.01 15.6 70.8 trace 2.1 0.92 Argus brown Farmer “P,” Putaruru
D/846 2.48 13.16 13.3 70.6 trace 5.4 1.17 Argus brown Dairy Factory, Tokoroa
Average: 2.72 12.59 13.3 72.0 0.05 3.6 0.95
– 195 –

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

D/446 2.77 13.50 67.6 0.27 1.6 0.66 Sudan brown Representative sample early grinding Puhipuhi.
D/832 2.20 14.62 16.8 66.6 trace 2.4 0.54 Between Argus and Brussels brown Untested in field. Puhipuhi.
D/844 3.20 15.03 17.2 71.2 trace 0.4 0.52 Brussels brown Tokaroa factory. Reputed good results cattle. Puhipuhi.
A/1115 9.59 13.01 16.6 62.3 0.28 0.28 1.3 2.86 Between Brussels and Antique brown Farmer “H,” Atiamuri.
D/837 7.96 14.40 16.2 71.8 0.29 1.2 2.20 Between Mars yellow and Antique brown Farmer “N,” Tokoroa.
D/840 9.53 14.34 16.9 68.2 0.28 0.92 2.54 Antique brown Factory, Tokoroa
D/842 7.52 14.33 16.3 72.3 0.33 0.6 2.22 Antique brown Farmer “N,” Tokoroa.
– 196 –

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

D/843 8.20 14.38 16.8 70.1 1.5 0.6 3.26 Raw sienna Farmer “H,” Atiamuri Ruatangata Iimonite. (Crushed and screened.) Good.
D/845 7.85 12.76 15.0 71.6 0.31 0.4 2.10 Between Antique and Brussels brown Farmer “P,” Putaruru
Average: 8.44 13.87 16.3 69.4 0.52 0.8 2.53
D/724 7.32 13.63 16.6 68.2 036 trace 3.06 Antique brown Representative sample.
D/776 5.89 13.83 15.7 69.8 0.33 1.8 1.72 Argus brown Rustangata heated. Untested in field.
D/736 3.21 11.63 11.1 71.4 traee 0.1 3.44 Antique brown Ruatangata ground by Fertilizer Co.
C/919 1.86 10.16 13.4 63.0 0.10 0.91 1.14 Between Brussels and Argus brown Onekaka. Untested in field.
D/777 5.60 14.50 18.2 64.1 0.17 0.4 1.41 Brussels brown Okaihau. Untested in field.
D/980 5.39 13.31 16.2 66.8 0.32 1.0 1.07 Between Brussels brown and Aigus brown Okaihau. Untested in field.
– 197 –

A number of commercial samples of “limonite” having known histories in the treatment of bush sickness were therefore collected from farms by Mr B. C. Aston and Mr C. R. Taylor, Analyst's assistant at Rotorua, and together with specimens of the original ores, were submitttd to partial analysis and to a number of special tests in an attempt to correlate nutritional efficacy with physical and chemical characteristics, and especially with the rate of solubility under conditions resembling those which might be met with in the ruminant digestive system.

The preliminary data obtained from a general survey of the samples are shown in Table I. In cases where several samples of the same or closely similar material were examined with concordant results, representative analyses only are given. For some of these samples no definite field reports are available; these, however, have been included, as they indicate the general uniformity of the samples received from the same source. From the groups indicated by the preliminary data, typical samples were selected for further study, the results being shown in the subsequent tables.

It will be seen that in the majority of cases where a sample had proved ineffective it was found to have both a low combined water and a high calcium carbonate content.

In the case of the commercial Onekaka sample with low combined water and low calcium carbonate content, no feeding test with unheated material has been made.

It is evident that if the rate at which the dilute gastric hydrochloric acid came into contact with the solid particles of lick were less than the rate at which it could dissolve the calcium carbonate, the iron oxide would probably be attacked only very slowly until all the lime had gone into solution. It is interesting to note in this connection the great difference in conditions obtaining in the ruminant and non ruminant stomachs respectively. In the ruminant the pretreatment of the vegetable food in the rumen together with large quantities of saliva results in the production of a considerable amount of alkali carbonates. On passage of the liquid material into the fourth stomach a large quantity of gastric hydrochloric acid must be used up in neutralising this carbonate before the contents can become acid. In an actual case of a sheep still warm (killed less than 1 hour) the rumen was filled with about 2 litres of green liquid (finely divided grass) having an alkalinity equal to N/8. The fourth stomach contained about 500 cc. of liquid, also strongly alkaline, so that apparently some time elapses before neutralisation takes place. In the contents of a second stomach examined, an acidity of N/30 was found. As the gastric juice has only about the same normality, it would seem that it must eventually be considerably more diluted than in the case of the non-ruminant where only a slight alkalinity due to the saliva has to be neutralised. It is worth

– 198 –

noting this point when considering the fact that ruminant animals are much more susceptible to bush sickness or iron starvation than non-ruminants.

In some experiments in which the quantities of dilute (0.36 per cent.) HCl and of limonite were so arranged that the calcium carbonate present materially reduced the concentration of acid, an inverse relationship was found between the amount of iron dissolved and the percentage of CaCo3. When the total quantity of the same strength HCl was so increased that the effect became negligible the amount of iron dissolved showed no correlation either with the field evidence or the analytical data.

A case has recently come under notice where Ruatangata limonite only was used for sheep and cattle on a Tokoroa farm. After initial good results, a considerable number of sheep commenced to lose condition. It was then found that carbonate of lime had been added to the limonite with the idea of improving the lick. On omitting the lime good results were once more obtained with the sheep. Cattle were not affected.

While, however, admixture of carbonate of lime may quite well be one factor in rendering “limonite” unassimilable, it is not the only one, as is shown by the case of the Onekaka ore * (heated).

Results by the “available iron” oxalic acid method suggested by Seelye and adopted by Rigg and co-workers (1932) showed relatively high solubility for all samples, and the differences between those proved to be effective and those proved to be ineffective did not seem to have great significance (Table II).

Table II.
Relative Solubilities of Limonite Samples in N/10 Oxalic Acid.*
Lab. No. % Fe2O3 Remarks.
A/1115 11.4 Ruatangata, coarsely screened, good results
D/742 11.5 " very finely ground
D/725 13.0 " representative sample
D/736 11.9 " ground in cement mill, heated
D/771 7.8 Puhipuhi, ground in cement mill, ineffective Farmer “H,” Atiamuri
D/773 8.6 " ditto. Farmer “P,” Putaruru
D/774 9.8 " ditto. Farmer “N,” Tokoroa

As it did not seem likely from these results that the differences in the availability of the various samples of limonite could be correlated with their solubility simply under varying conditions of

[Footnote] * Since this paper was written, an article by Rigg and Askew has appeared in Empire Journal of Experimental Agriculture, Vol. 11, No. 5, January, 1934, on “Soil and Mineral Supplements in the Treatment of Bush Sickness.” In a course of feeding experiments with sheep on bush sick pasture at Glenhope, it was found that good results were obtained with drenches of Nelson garden soil and iron ammonium citrate, whereas typical bush sickness developed in control group and the group drenched with Onekaka limonite. The conclusion is reached that the failure of Onekaka limonite to overcome bush sickness shows that “the supply solely of iron containing compounds is not sufficient for the prevention of ailment.” At the same time, it is stated that the (efficacious) iron ammonium citrate used for drenching was chemically pure.

[Footnote] * Analysis by Mr D. F. Waters.

– 199 –

acidity, it was thought that some other factor, such as rate of solution rather than solubility at equilibrium, or susceptibility to reducing agents, might be concerned.

In some preliminary experiments it was found that the stomach semi-liquid contents of sheep had a reducing action equivalent to between N/100 and N/200 iodine. This is a matter of importance, since Johnson (1924) has shown that ferrous oxide is much more soluble in dilute hydrochloric acid than ferric oxide. Moreover, in the present investigation we have found that the addition of a mild reducing agent to dilute hydrochloric acid greatly enhanced the solubility of the limonite, while in a buffered acetic acid solution (pH 4.0) in which the limonite was practically insoluble the addition of sodium hydrosulphite not only rendered the limonite readily soluble, but within limits the amount of iron dissolved was directly proportional to the amount of reducing agent added (see Table III and graph). It was thought that further information on the relative susceptibilities of the limonites to reducing agents might explain to some extent the wide differences in the efficacy of the limonites as shown in the field tests. Three series of experiments were therefore devised to evaluate the solubilities of the limonites, using as reagents sugar-hydrochloric acid, acetic-sodium hydrosulphite, and hydrogen sulphide solutions respectively.

(a) Solubility of Limonite in Sugar-Hydrochloric Acid.

The reagent for this experiment was prepared by dissolving 100 grams of commercial sucrose in one litre of N/10 hydrochloric acid made up from freshly boiled distilled water. In order to facilitate inversion the sugar was added while the solution was still hot. In the solubility test 0.5 gram of limonite was placed in a 200 cc. conical flask with 50 cc. of the reagent and immediately stoppered with a tight fitting cork provided with a bunsen valve to protect the solution fom oxidation by air. After shaking the flask in boiling water for a period of 15 minutes, the solution was quickly cooled, made up to 100 cc. with distilled water, and immediately filtered through a No. 42 Whatman paper. After 20 cc. of filtrate had been collected the funnels and filter papers were removed to exclude the possibility of further solution of the limonite, and a suitable aliquot taken for the iron estimation. It was found that by keeping strictly to the above conditions duplicate samples yielded concordant results.

On reference to Table I it will be seen that all the efficacious samples (as judged from the yield of milk in the case of cows or from increase of weight in the case of sheep) of manufactured Ruatangata limonite had a solubility several times greater than the average of the relatively ineffective manufactured Puhipuhi limonite samples. The Onekaka (heated) sample also had a low solubility. It was found, however, that the Puhipuhi ore as mined had a low solubility as did an early sample of ground and apparently unheated material from the same locality which was reported to have given good results at

– 200 –

Tokoroa on cattle. Low figures were also obtained for the manufactured Okaihau product. A sample of the iron oxide (haematite) used successfully with cattle in Kenya Colony gave a very low solubility.

(b) The Solubility of Limonite in Acetic Acid-Sodium Acetate Solution containing Sodium Hydrosulphite as a Reducing Agent.

Elvehjem, Hart and Sherman (1933) have devised a chemical method for determining the biologically available iron of foods. Their method depends on the fact that under specified conditions an acetic acid-sodium acetate buffer solution containing sodium hydrosulphite will dissolve out “available” iron, while the iron bound in complex form such as haematin iron which is unavailable for haemoglobin formation is not attacked by this reagent. The chemical method yielded in the hands of these workers results in good agreement with those obtained from animal feeding experiments. It was hoped that this method would give a measure of the biological efficiency of the various limonites. Actually, as shown in Table III, when the limonites are treated according to the Elvehjem, Hart and Sherman method no differences between the various limonites were found, each limonite having the same amount of soluble iron in equilibrium with the reagent.

As the total solubility did not discriminate between the different limonites it was decided to measure the rate of solubility according to the following method. 0.25 gram of limonite was weighed into a test-tube with 0.25 gram of sodium hydrosulphite. 20 cc. of aceticsodium acetate solution (buffered to pH 4.0) were rapidly transferred from a measuring cylinder into the test-tube and the time noted. The test-tube was then corked as rapidly as possible and vigorously shaken until 10 seconds before the appointed time, when the solution was filtered through a No. 42 Whatman paper, the filtrate being removed 10 seconds after the appointed time. In the case of the 15 seconds period the filtration proceeded for only 10 seconds, so that the aliquot available for analysis amounted to less than 1 cc., necessitating the use of a serum pipette. For the other time intervals a 1 cc. aliquot was taken for the estimation. The iron estimation was made according to Hill's (1931) α α ′ dipyridyl method. The method as originally outlined was not suitable for the present investigation and certain modifications were introduced. It was found that the iron dipyridyl colour could be readily measured on the Lovibond tintometer. Measured volumes of standard 2M/10,000 ferrous ammonium sulphate were added to a measured excess of the 12M/10,000 standard α α ′ dipyridyl solution so as to produce concentrations of iron varying from 0.0010 to 0.0060 mgm. of iron per cc. On plotting the concentration of iron against the Lovibond red units a straight line was obtained as shown below in Graph I.

– 201 –

Graph I.

Picture icon

Mgms. Fe per cc.

The red colour within the limits of the concentrations measured is thus directly proportional to the concentration of the iron in solution.

Table III.
The Effect of Sodium Hydrosulphite on the Solubility of Limonite in an Acetic Acid-Sodium Acetate Buffer Solution (ph 4.0).
% Iron Dissolved (as Fe2O3) at Equilibrium.
Lab. No. No reducing agent. 1.25% reducing agent. 2.5% reducing agent. 3.75% reducing agent.
A/1115 (Ruatangata) 0.0 22.4 36.5 52.5
D/725 (Ruatangata) 0.0 23.1 41.2 52.5
D/773 (Puhipuhi) 0.0 22.3 41.2 50.4
D/844 (Puhipuhi) 0.0 22.9 34.3 54.9
C/919 (Onekaka) 0.0 23.4 36.5 sample all used.
– 202 –
Picture icon

% Fe2O3 Dissolved

Table IV.
The Rate of Solubility of Limonite in a ph 4.0 Buffer Solution containing 1.25% Sodium Hydrosulphite.
Lab. No. 15 secs. 30 secs. 60 secs. 90 secs. 5 mins. 1 hour.
A/1115 (Ruatangata) 5.1 8.9 14.7 18.3 22.4 22.3
D/724 (Ruatangata) 7.8 10.3 20.8 21.8 24.5 23.1
D/773 (Puhipuhi) 2.1 5.5 11.2 13.0 20.2 22.3
D/844 (Puhipuhi) 1.6 3.9 7.8 9.2 22.8 22.8
C/919 (Onekaka) 1.4 3.7 9.8 sample all used. 23.4 23.4

It will be observed from the above table that although all the limonites reached equilibrium with the solvent in approximately five minutes they approach equilibrium at a different rate. The discriminating factor between the various limonites does not therefore depend on the solubility at equilibrium, but on the rate of solution. Under these circumstances the most suitable period of solution for a comparison of limonites is the shortest period which can be accurately and conveniently measured. It was found that a 30 second period gave good agreement between duplicate determinations and could be adopted as a useful standard for comparison. It is interesting to note that the hydrosulphite method yields similar differences between the various limonites to those given by sugar-hydrochloric acid method.

– 203 –

(c) The Effect of Hydrogen Sulphide on the Solubility of Limonite in Hydrochloric Acid.

It is well known that when hydrated iron oxide is treated with hydrogen sulphide in neutral solution it becomes dark coloured, H2S being removed from solution to form iron sulphide. According to Thorpe's Dictionary of Applied Chemistry (1921 edition) the reaction Fe2O3 + xH2O + 3H2S = 2 FeS + S + (x + 3) H2O. takes place in neutral or faintly acid solutions, while in weakly alkaline solutions the reaction is Fe2O3 + xH2O + 3H2S = Fe2S3 + (x + 3) H2O. In strongly acid or alkaline solution the reactions are inhibited. With natural hydrated iron oxides the extent to which the reactions proceed varies enormously. The highly hydrated earthy samples take up large quantities of H2S while the hard, less hydrated ores approaching haematite in composition are much less affected.

In preliminary experiments the various materials used in the stock-feeding trials gave markedly different results when tested for their capacity for absorbing H2S, and after such treatment similar differences were apparent in their solubility in dilute (0.36 per cent.) hydrochloric acid. It was found that if the acid and H2S were added simultaneously very little iron was dissolved. The colours given by the fine suspensions of the limonites after treatment with H2S, when examined by reflected light, afford an interesting method of differentiating between the various deposits, the original colour of the sample (as affected by heating in the cement mill, or by fineness of grinding) appearing to make little difference.

The following method was adopted for estimating the relative effect of H2S on the solubility of limonites in dilute HCl:—0.5 gm. limonite is shaken for 30 minutes with 80 cc. N/20 freshly prepared H2S solution in a closely stoppered flask or test-tube of 100 cc. capacity. 10 cc. of N, HCl are then added and the mixture shaken for a further 10 minutes. It is then filtered through a dry paper and the iron determined on a suitable aliquot. Duplicates were found to agree closely. Some typical results are given in Table V.

Table V.
Effect of Treatment with h2s on the Solubility of Lemonite.

Lab. No. Colour of H2S Suspension. % Fe2O3 Dissolved in 0.40 Hcl. Remarks.
D/771 grey (purplish) 7.7 Puhipuhi. Relatively ineffective.
D/772 grey 5.9 " "
D/774 grey 5.0 " "
D/838 black (purplish) 8.4 " "
D/446 grey 4.8 " Early grinding.
grey (purplish) 5.2 " Ore.
A/1115 purple (bright) 16.7 Ruatangata. Good.
D/725 " " 22.0 " "
D/698 " " 22.8 " "
D/810 " " 20.9 " Ore.
D/742 navy blue 22.0 Ground by Fertilizer Company.
D/776 purple 20.2 " " "
D/1502 purple (dull) 5.8 Onekaka. Relatively ineffective.
D/1007 bright red 0.5 Haematite. Kenya Colony.
D/1468 purple 31.9 “Lux” imported for gas purification.
– 204 –

It will be seen that the solubilities given by this method follow the same general order as with methods A and B, but the discrimination between the samples is wider.

Discussion of Solubility Results obtained by Methods A, B and C.

Consideration of the results under sections A, B and C suggests strongly that the differences in efficacy between the various types of iron oxide used in feeding trials are a matter of degree only. As there are also degrees of bush sickness, both in the capacity of various types of soil for producing the disease, and in the response of the different classes of livestock on any one deficient area, it is evident that close correlation of field and laboratory data cannot be expected except in cases where feeding experiments with the different products have been carried out in one locality using similar minerals.

There is evidence from experiments conducted by the Rowett Institute on Nakuruitis (Orr and Holm 1931) (a deficiency disease closely related to bush sickness and affecting cattle on the volcanic soil of Nakuru in Kenya Colony) that the quantity of iron oxide consumed is a very important factor in the prevention and cure of the disease. The control and low iron diet groups showed all the typical symptoms of Nakuruitis. The third group were in normal condition, while the fourth group, which received approximately two and a-half times the amount of iron consumed by the third group, were in prime condition, having gained more than double the weight of that group.

In the case of enzootic marasmus, an iron deficiency disease possibly identical with bush sickness and affecting the sheep and cattle of the Denmark district in Western Australia, Filmer (1933) has shown that the curative dose of iron may vary enormously according to the iron compound used. The effective dose for sheep was found to vary from .02 gram to 11 grams per head per day for dried liver extract and limonite respectively. In the experiments with ferrous carbonate it was shown that the availability of the iron differed considerably according to the source of the material. Whereas daily doses of iron carbonate in the form of spathic iron ore containing as much as 2.8 grams of iron were ineffective in combating the disease, Blaud's pills also containing iron as ferrous carbonate were effective in daily doses corresponding to 1.16 grams of iron. Experiments on calves with various iron compounds gave results similar to those obtained for the sheep.

As the bulk of the iron is excreted in the faeces it is difficult to account for the greater efficacy of large amounts of iron unless a “large head of iron” in the intestine is necessary to secure the absorption of the relatively small amounts required for haemoglobin formation, Morris (1933). In considering the therapeutic value of the dried liver preparation Filmer has estimated that the addition of iron present in such a preparation would in the case of his experiments amount to the addition of one-twentieth of the iron normally present in the unsound pasture and it therefore seems unlikely that the iron content of the dried liver is the major curative factor. To explain this anomaly Filmer has put forward the hypothesis that

– 205 –

the effectiveness of the various iron compounds depends on the presence of some unknown mineral necessary for the metabolism of iron. In our opinion such a hypothesis does not seem to be justified; for pure iron compounds such as iron and ammonium citrate effectively cure enzootic marasmus as well as bush sickness. According to Reakes and Aston (1919) pure iron tartrate and iron acetate * are also efficacious in curing bush sickness. Moreover, a fusion analysis of the effective Ruatangata limonite used in our experiments did not reveal the presence of any element which is not normally present in the pasture.

In view of the greater effectiveness of ferrous iron as compared with ferric iron as a cure for certain forms of anaemia, Morris (1933), and in the light of our present investigations which show limonites to be much more readily soluble in acid solution in the presence of reducing agents, we would suggest the possibility that the effectiveness of liver preparations as a cure for enzootic marasmus is due to the presence of reducing substances such as glutathione normally present in the liver enhancing the solubility of the otherwise unavailable iron present in the unsound pasture. Such a suggestion might also explain the enhancing effect of the dried stomach preparations when used in conjunction with iron and ammonium citrate and the fact that dried pig's stomach alone gave no response when administered to lambs suffering from enzootic marasmus. The ineffectiveness of liver ash can be similarly interpreted.

Although no field tests have been made in New Zealand to ascertain the minimum quantity of limonite necessary to maintain the health of the stock, it appears from the report of a sheep farmer at Puketurua, who obtained excellent results when his stock were fed with the Ruatangata limonite, but poor effects with the Puhipuhi limonite when fed at the same rate, that the severity of the bush sickness was checked to some extent by increasing the inferior limonite in the salt lick from 50 per cent. to 80 per cent.

In conclusion it thus seems possible that there is a critical value in the quantity of iron necessary to keep the stock healthy. This value would naturally vary with the nature of the limonite and with the ruminant under consideration. Thus sheep appear to be more susceptible to bush sickness than cows.

The Effect of Heating on the Physical and Chemical Properties of Limonite.

Apart from the earlier work of Fisher (1910) and the later and more extensive investigations of Posnjak and Merwin (1919) on the dehydration of limonite and other hydrous hydrated iron oxides, there appears to have been no precise record of the physical and chemical changes which take place when these oxides are heated.

[Footnote] * Personal communication.

– 206 –

Fisher has shown that in the case of limonite, water is lost rapidly below 100° and then sparingly although continuously up to 165°, at which temperature the monohydrate appears to be stable. Above this temperature the yellow monohydrate decomposes to form red haematite. Recent investigations, however, suggest that a definite transition temperature of the monohydrate to the anhydrous oxide has not been established.

Actually the detection of ferric oxide hydrates is complicated by the fact that the hydration of ferric oxide is irreversible, preventing the application of usual methods such as the measurement of the vapour pressure of the hydrate at constant temperature where it can be shown by the phase rule that the composition of each new hydrate will be accompanied by a sudden decrease in vapour pressure. In order to avoid this difficulty Posnjak and Merwin applied the LeChatelier method for irreversibly hydrated substances to the study of limonite and related iron ores. The method consists essentially in measuring the temperature at definite intervals of time while the substance is heated at a uniform rate. If no chemical or physical change occurs, a straight line graph is obtained on plotting the temperature against time. If, however, a chemical or physical change occurs, it is accompanied by the absorption or evolution of heat, and in place of a straight line, a curve (the sharpness of which depends on the stability of the component or components concerned in the transition) is shown on the graph at the decomposition temperature. The hydrous hydrated oxides of iron were shown by this method to have a decomposition temperature of approximately 300°. This figure is considerably higher than obtained by Fisher from simple dehydration. Posnjak and Merwin attribute the difference to the slow rate of reaction and physical condition of the limonite.

Posnjak and Merwin also studied the dehydration by heating the minerals to constant weight at fixed temperatures. The curves obtained by plotting the percentage of water against the temperature were found to consist of three distinct parts, the middle portion of which was considered to be due to the decomposition of the monohydrate, while the upper and lower portions of the curve probably corresponded to the capillary and adsorbed water.

It is well known that hydrated iron oxides on strong ignition become much less soluble in concentrated hydrochloric acid, but so far as the authors are aware no quantitative measurements have been made, and in view of the fact that certain ineffective limonites were known to have been heated, it was decided to make a systematic survey of the changes in the solubility, specific gravity, and hydration of representative samples when heated to a definite temperature for a known period of time. The results of this investigation are presented in the table below.

– 207 –

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

Table VI.
The Effect of Heating on the Chemical and Physical Properties of Limonite.
Sample. Solubility in sugar-hydrochloric acid Solubility. ratio heated Unheated sample. Specific gravity at 20°. % 11, O (free and combined). Remarks.
A/1115— as received 2.86 2.80 22.0 Good limonite from Ruatangata deposits. Use by Farmer “H” at Atiamuri.
" heated to 100° for 24 hours. 2.92 1.02 2.95 12.4
" heated to 184° for 1.¼ hours. 3.74 1.31 3.26 10.8
" heated to 290° for 1.½ hours. 3.00 1.05 3.61 1.8
" heated on blowpipe for 10 mins. 0.08 0.028 4.16 0.0
D/724— as received 3.06 2.97 21.0 Representative sample from Ruatangata deposits.
" heated to 100° for 24 hours. 3.06 1.00 3.10 13.6
" heated to 184° for 1.½ hours. 3.76 1.23 3.39 11.6
" heated to 290° for 1.½ hours. 3.06 1.00 3.95 3.4
" heated on blowpipe for 10 mins. 0.09 0.029 4.47 0.0
D/773— as received 1.14 3.58 10.5 Ineffective limonite from Puhipuhi deposits. Used by Farmer “P,” Putaruru.
" heated to 100° for 24 hours. 3.67 8.8
heated to 184° for 1.½ hours. 1.14 1.00 3.86 8.3
" heated to 290° for 1.½ hours. 2.08 1.83 4.03 3.0
" heated on blowpipe for 10 mins. 0.68 0.60 4.37 0.0
D/446— as received 0.66 3.42 15.6 Representative sample from Puhipuhi deposit.
" heated to 100° for 24 hours. 0.66 1.00 3.46 12.8
" heated to 184° for 1.½ hours. 0.79 1.19 3.84 12.1
" heated to 290° for 1.½ hours. 0.69 1.05 3.95 5.7
" heated on blowpipe for 10 mins. 0.29 0.44 4.47 0.0
– 208 –

The solubility of the limonites was determined according to the usual procedure except that an allowance was made for the loss of water in the heated samples by taking a weight corresponding to the same amount of ferric oxide as was present in 0.5 grams of unheated sample. In each series of solubility determinations on the heated samples the corresponding unheated samples were used as controls, thereby allowing for any difference in conditions between successive experiments. The heating at 100° was done in an electric oven, while the intermediate temperatures at 184° and 290° were obtained by permitting aniline and glycerine vapour respectively to flow around the outer surface of a test-tube containing the sample together with a small wire stirrer. As preliminary heating of the limonite in boiling distilled water or evacuation of the limonite under reduced pressure made little or no change in the observed specific gravity, the specific gravity determinations were made directly in a specific gravity bottle according to the usual method.

The samples presented in Table VI are representative of the Ruatangata and Puhipuhi deposits respectively. Sample D/773, however, has very probably been heated as is shown from its colour and low percentage of combined water. This suggestion is also supported by the small loss of water and absence of change in solubility on heating to 184° for 1.½ hours. It will be observed from the table that on heating the limonites to 100° for 24 hours little or no change occurs beyond a loss of free water accompanied by a small rise in the specific gravity. With the exception of sample D/773, at 184° a further small amount of free water is lost accompanied by a disproportionate increase in specific gravity and a significant rise in solubility, suggesting an alteration of the physical or chemical nature of the limonite. On heating to 290° most of the remaining water is lost, and the solubility returns to that of the original sample except in the case of D/773, where the solubility is considerably increased. At red heat the remaining water is lost and the oxides tend to the same specific gravity while their solubilities are considerably diminished. The diminution of solubility commenced at between 500° and 600° C. in the electric muffle, and its extent is evidently characteristic of a particular ore being much more marked in the case of A/1115 than for D/446. It should also be mentioned that further heating of the iron oxide brought about no further change of solubility.

As an additional check on the source of the limonites, ores taken from the Ruatangata and Puhipuhi, Okaihau and Onekaka deposits respectively were ground in the laboratory so that 80 per cent. passed 120 mesh sieve, and analysed with the following results:—

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

Deposit. Specific Gravity. Loss on Ignition. Moisture. Solubility in sugar-hydrochloric acid.
Puhipuhi deposit 3.51 12.58 2.55 0.52
Ruatangata deposit 3.08 12.45 6.22 4.11
Okaihau deposit 3.19 14.07 5.87 1.74
Onekaka deposit 3.44 11.14 1.79 0.48
– 209 –
Table VII.
The Effect of Fineness of Grinding on the Solubility of Limonite in Sugar-hydrochloric Acid.
Fineness of Grinding. A/1115—From Ruatangata deposits. Gave good results. Farmer “H,” Atiamuri. D/724—Representative ample from Ruatangata deposits. D/771—From Puhipuhi deposits. Farmer “H,” Atiamuri. D/446—Representative sample from Puhipuhi deposits. D/366—Sample from Okaihau deposits.
Retained on 30 mesh 2.0 trace trace trace trace
Retained on 60 mesh 29.2 0.5 trace trace 11.0
Retained on 90 mesh 19.4 9.0 2.0 2.0 24.0
Retained on 120 mesh 5.2 9.0 3.5 4.0 9.0
Retained on 150 mesh 5.0 11.0 7.5 5.0 9.0
Retained on 200 mesh 6.0 9.0 6.0 6.0 9.0
Passed 200 mesh 33.2 61.5 81.0 83.0 38.0
Solubility in Sugar-Hydrochloric Acid 2.86 3.06 0.94 0.66 1.01
Colour Between Brussels brown and antique brown. Antique brown. Between argus brown and antique brown. Sudan brown. Brussels brown.

Fineness of grinding tests were made on several representative samples in order to correlate the fineness with solubility in sugar hydrochloric acid and with the effectiveness of the limonite as judged from field tests. In the case of a material composed of solid grains reduction of the grains by grinding would expose new surfaces, thereby increasing the solubility. If, however, the grains were composed of minute particles loosely held together so as to form porous aggregates, then further grinding within certain limits should not expose fresh surfaces and so increase the solubility. Actually, as shown in the table, the rate of solubility of the limonites depends mainly on the source of the limonite and not on the state of fineness of the particles as determined by sieving. In this connection it should be emphasized that from the point of view of palatability the limonite used for stock feeding purposes must be finely ground, as sheep will not eat gritty material. In order to test the homogeneity of the limonite, a coarse fraction retained on a 50 mesh sieve was ground to pass a 200 mesh sieve. Solubility tests showed no appreciable difference between this finely ground material and the material originally passing the 200 mesh sieve. A careful examination under the microscope of the Ruatangata and Puhipuhi limonites showed that few of the particles were massive and most possessed a porous structure, being channelled and perforated with small holes. This condition was particularly marked in the case of the Ruatangata limonite. In conclusion, it would appear that from the evidence so far available the structure of the limonite is essentially that of a porous material composed of very fine homogeneous particles.

– 210 –


Hydrated iron oxides have very ill defined compositions, and in any specimen of natural ore it is difficult to determine what iron compounds are present.

In the natural deposits used for stock feeding the solubility of the iron in dilute (0.36 per cent.) hydrochloric acid was very small and irregular. When reducing'agents were added the solubility was much increased, the solubility at equilibrium (in the case where hydrosulphite was used) being similar for all samples and proportional to the amount of reducing agent present.

When, however, the rate of solution in the presence of reducing agents was determined it was found to show considerable differences between the samples from different sources, samples definitely known to be highly effective in curing bush sickness yielding high figures and ineffective samples low figures. Some apparent anomalies require further field experiments for elucidation. After treatment with hydrogen sulphide, great differences were found in the solubilities of the limonites in dilute hydrochloric acid. As the H2S could only presumably attack the surface of the particles, it appears that the chief differences between the samples are to be sought in the physical state of the finest particles constituting the aggregates. In the good samples these particles would appear to be colloidal and loosely bound together, while in the inefficient samples they are evidently bound more securely as shown from specific gravity determinations. Hence in the earthy Ruatangata material the degree of fineness of grinding, or a considerable amount of heating (complete dehydration at 500° C.) have little effect on the solubility, while in the case of the harder Puhipuhi deposit, finely ground or heated samples appear in several cases to have higher solubilities than the original ore.

According to Hawke (1926), powerful reducing reactions take place in the intestine, especially in its lower portion, and these can be measured by the amount of ferrous hydroxide produced from ferric hydroxide after passage of a test meal. Paton and Orr (1920) consider the large intestine to be the seat of important digestive and absorbtive processes in herbivores so that considering that ferrous hydroxide is slightly soluble in neutral and alkaline solutions it is feasible that some absorption of iron may take place after reduction in the lower part of the intestines. Recently Lintzel (1931) from his experiments on white rats has stressed the importance of the formation of ferrous iron before the iron can be assimilated. It will be seen therefore that ease of reducibility may be an important factor in availability to the animal of the iron in a “limonite.”

Hawke states that minute solid particles of iron compounds (hydroxides) are ingested and may be carried by white blood corpuscles through the intestine membranes into the lymph. Abderhalden (1911) also mentions the same phenomenon. It is evident that for this type of assimilation only the most minute colloidal particles could be utilised and such fineness could not be attained by ordinary grinding, but would have to be a property of the ultimate constituent particles of the original deposits.

– 211 –

In conclusion it should be emphasized that the present investigation is intended only to give a chemical picture of the problem with a view to facilitating the future identification of limonites. The correlation of the chemical properties with the efficacy of the sample in field experiments in the present work should also enable the chemist to suggest whether or not a given limonite is suitable for stock feeding purposes. Metabolism experiments are not only expensive, but in the case of limonite, would probably give inconclusive results owing to the difficulty of discriminating between the absorbed and excreted iron and to the absence of definite knowledge of the processes of iron assimilation in the digestive system of ruminants. It is considered, however, that a thorough survey of the digestive system of the sheep for soluble iron after feeding with limonite would throw considerable light on the nature of iron assimilation and perhaps decide the relative efficacy of a given limonite for stock feeding.


The authors wish to express their thanks to Mr B. C. Aston, Chief Chemist, Department of Agriculture, for information, suggestions, and for permission to carry out the work. Thanks are also due to Messrs S. G. Brooker and I. G. McIntosh for assistance in the analysis of the samples, and to Mr D. H. Le Souef, Veterinarian, for obtaining fresh sheep's viscera.


  • (1) Fusion analyses of Ruatangata, Puhipuhi, and Onekaka ores show that in addition to iron oxide, aluminium, silicon, and water the following elements are present in some or all of the ores, in small amounts:—Titanium, calcium, magnesium, manganese, phosphorus, sulphur, barium, sodium and potassium, together with traces of copper, chromium, arsenic, antimony, nickel and cobalt. The differences in chemical composition between the various deposits would appear, however, totally inadequate to account for the variations observed in the feeding value of the ores.

  • (2) The analysis of 21 limonites representative of four deposits for moisture content, loss on ignition, ferrous and ferric oxide, calcium carbonate and colour estimation show that in the majority of cases the samples found by field trials to be ineffective in preventing bush sickness in stock are low in combined water and contain appreciable amounts of calcium carbonate.

  • (3) The solubilities of representative samples in N/10 hydrochloric acid or in N/10 oxalic acid show no striking correlation with the field evidence.

  • (4) Evidence is adduced from solubility experiments in sugar-hydrochloric acid solution, sodium acetate-acetic acid buffer solutions containing hydrosulphite, and hydrochloric acid solutions, after treatment with hydrogen sulphide that the limonites are much more soluble in the presence of reducing agents. The rate of solubility evaluated

– 212 –
  • from these reagents gave consistent results, the effective Ruatangata samples showing a much greater rate of solution than the ores from the other deposits.

  • (5) In the case of the sodium-acetate-acetic acid buffer solution ph 4.0 in which the limonite samples are insoluble the amount of iron dissolved at equilibrium is directly proportional to the amount of sodium hydrosulphite present and practically independent of the nature of the limonite.

  • (6) The evidence so far obtained suggests that the variations in the feeding value of the ores are to be correlated with the rate of solution of the iron in acid reducing agents rather than with the measurements of total solubility of the iron at equilibrium.

  • (6a) It is therefore suggested that the rate of solution under prescribed conditions of a given limonite may prove a laboratory method of determining the relative effectiveness of that limonite for curing bush sickness.

  • (7) A study has been made of the effect of heat on the chemical and physical properties of limonite. No striking change of solubility occurs unless the samples have been heated about 550°.

  • (8) Fineness of grinding as indicated by sieving tests does not appreciably influence the solubility of limonites in sugar hydrochloric acid solution.

  • (9) The evidence adduced from fineness of grinding measurements, microscopical examination and solubility data supports the idea that the limonites are composed of fine particles or colloids loosely bound together to form porous aggregates. Specific gravity determinations support the idea that these particles are more closely held together in the case of the inferior samples than in the case of the more effective samples.


Abderhalden, E. (Translation by Hall, W. T., and Defren, G.), 1911. Text Book of Physiological Chemistry, first edition, Wiley and Sons, New York, p. 390.

Clarke, F. W., 1924. The Data of Geochemistry, fifth edition, Government Printing Office, Washington, p. 534.

Elvehjem, C. A., Hart, E. B., and Sherman, W. C., 1933. The Availability of Iron from Different Sources for Haemoglobin Formation, Journ. Bio. Chem., Vol. 103, p. 61.

Filmer, J. F., 1933. Enzootic Marasmus of Cattle and Sheep, Australian Veterinary Journal, Vol. 9, pp. 163–179.

Fischer, H. W., 1910. Zs. Anorg. Chem., Vol. 66, p. 37. Quoted by Posnjak and Merwin (loc. cit.).

Friend, J. N., 1921. Text Book of Inorganic Chemistry, Charles Griffin and Co., Ltd., London, Vol. 9, Part 2, pp. 9–33.

Hawke, P. B., and Bergeim, O., 1926. Practical Physiological Chemistry, ninth edition, J. and A. Churchill, London, p. 283.

Hell, R., 1931. A Method for the Estimation of Iron in Biological Material, Proc. Roy. Soc. of London, Series B, Vol. 107. pp. 205–214.

– 213 –

Johnson, M. O., 1924. Manganese Chlorosis of Pineapples, U.S. Dept. of Agriculture Bull. No. 52, p. 25.

Lintzel, W., 1931. Absorption of Food Iron as Ferrous Iron, Biochem. Ztschr., Vol. 263, pp. 173–186. Nutrition Abstracts and Reviews, 1934, Vol. 3, p. 786.

Morris, N., 1933. A Review of Observations on Diet in Relation to Health and Disease, Nutrition Abstracts and Reviews, Vol. 2, p. 667.

Orr, J. B., and Holm, A., 1931. Report, Sixth Report of the Economic Advisory Council Committee on the Mineral Content of Natural Pastures, His Majesty's Stationery Office, London, pp. 43–48.

Paton, D. N., and Orr, J. B., 1920. Essentials of Physiology for Veterinary Students, third edition, Green and Son, Edinburgh, pp. 337–346.

Posnjak, E., and Merwin, H. E., 1919. The Hydrated Ferric Oxides, Amer. Jour. Sci., Vol. 47, pp. 311–348.

Reakes, C. J., and Aston, B. C., 1919. Curative Treatment of Bush Sickness by Iron Salts, N.Z. Jour. Agric., Vol. 18, pp. 193–197.

Rigg, T., and Askew, H. O., 1932. Bush Sickness, part III, N.Z. Dept. of Sci. and Ind. Res. Bull., No. 32, p. 58.

Weiser, H. B., 1926. The Hydrous Oxides, first edition, McGraw-Hill Book Company, Inc., New York, pp. 34–38.