The usefulness of soil phosphorus (P) tests used in routine soil analyses is limited by the fact that a single measurement cannot encompass all P-related factors potentially affecting plant performance. In this work, we performed micropot (15 mL) experiments to test the hypothesis that the predictive value of two common soil P tests (Olsen P and CaCl 2 -P) can be improved by considering properties commonly measured in laboratory analyses. Forty-nine sets of soils ranging widely in properties were used for this purpose, each set consisting of samples with similar properties but differing in P status. Ryegrass and turnip were grown in a chamber for 30 days in two separate experiments and their yields at harvest recorded. The critical Olsen P and CaCl 2 -P levels, which were taken to be those corresponding to 95% asymptotic yield as calculated from data fitted to a Mitscherlich equation, were greater for turnip than for ryegrass, probably as a result of the difference in yield (49 and 160 mg dry matter/micropot on average for ryegrass and turnip, respectively) and hence in P requirements between the two species. Critical Olsen P spanned narrower ranges than critical CaCl 2 -P in both crops and is therefore seemingly the more robust of the two tests. Both critical P values exhibited moderate correlations with soil properties. Thus, critical Olsen P was (a) lower in soils with a medium pHwhich is consistent with the fact that the bicarbonate solution method tends to overestimate plant-available P in strongly acid and calcareous soils; (b) positively correlated with pH and carbonate content in calcareous soils; and (c) uncorrelated with soil properties in noncalcareous soils. On the other hand, critical CaCl 2 -P in some soil groups was negatively correlated with some properties increasing the P buffering capacity of soil (e.g. Fe oxide content). Taken together, our results suggest that routinely measured soil properties help to predict critical Olsen P better than critical CaCl 2 -P.

Previous pot cropping and laboratory incubation experiments were consistent with field observations showing that temporary flooding before cropping can increase the availability of soil Fe to plants. To study the effect of temporary flooding on changes in soil Fe phytoavailability we used 24 highly calcareous, Fe chlorosis–inducing soils to carry out a pot experiment where peanut and chickpea were successively grown after flooding for 30 d. At the end of the cropping experiment, the preflooded soil samples exhibited higher concentrations of acid oxalate‐, citrate/ascorbate‐ and diethylenetriaminepentacetic acid (DTPA)–extractable Fe (Feox, Feca, and FeDTPA, respectively) than the control (nonflooded) samples. Also, Feox and Feca exhibited no change by effect of reflooding of the cropped soils or three wetting–drying cycles in freeze‐dried slurries of soils previously incubated anaerobically for several weeks. Leaf chlorophyll concentration (LCC) in both peanut and chickpea was greatly increased by preflooding. The best predictor for LCC was Feox, followed by Feca and FeDTPA. The LCC–soil Fe relationships found suggest that the Fe species extracted by oxalate and citrate/ascorbate from preflooded soils were more phytoavailable than those extracted from control soils. However, the increased phytoavailability of extractable Fe forms was seemingly limited to the first crop (peanut). Flooding dramatically increased FeDTPA; however, high FeDTPA levels did not result in high LCC values, particularly in the second crop. Therefore, this test is a poor predictor of the severity of Fe chlorosis in preflooded soils.

The aim of this paper was to compare the concentration of P in soil extracts prepared with water and a ‘soil solution proxy’ (‘SSP’, that is, a salt solution similar in ionic composition and strength to the actual soil solution) with that in 0.01 m CaCl2 extracts, which is usually taken as a measure of soil P intensity. Seventy widely ranging agricultural soils from the Mediterranean part of Spain were used. Soil/solution ratio was 1:10 and extraction time 3 days. For 0.01 m CaCl2, a short extraction time of 30 min was also used as the reference method. CaCl2‐P(3 days) and CaCl2‐P(30 min) were not significantly different for the 40 noncalcareous soils group, but CaCl2‐P(3 days) was significantly larger than CaCl2‐P(30 min) for the 30 calcareous soils group. The Water‐P/CaCl2‐P(30 min) ratio was not significantly related to any soil property, its mean being 6.3 for the noncalcareous and 5.8 for the calcareous soils group. The mean SSP‐P/CaCl2‐P(30 min) ratio was 2.6 for the noncalcareous and 3.1 for the calcareous soils group, and decreased slightly with increasing ionic strength of the soil solution in the noncalcareous soils group. These results are consistent with the promoting influence of the Ca ion and ionic strength on P adsorption by permanent‐charge soils. The fact that extraction with 0.01 m CaCl2 generally results in underestimation of the actual concentration of P in the soil solution should be considered when CaCl2‐P is used as a soil P test.

Little attention has so far been given to the relationship of Olsen P to solution P at P levels in the soil solution within the range relevant to plant growth. In this work, we investigated the relationship between these two P forms in 24 noncalcareous and 25 calcareous Spanish soils, using the concentration of P in the 0.01 M CaCl 2 extract (1:10 soil:solution ratio; 30 min shaking) as a proxy for the P concentration in the soil solution. Application of the Freundlich equation (y = ax b ) in linearized form provided a parsimonious, largely accurate description of the Olsen P/CaCl 2 -P relationship. The average value of exponent b, 0.53, suggests that Olsen P relates to the quantity of P in soil rather than to its intensity. The estimated Olsen P values at a CaCl 2 -P concentration of 0.01 and 0.03 mg/L (viz., Olsen P 0.01 and Olsen P 0.03 , respectively) varied over wide ranges (0.6-33.9 and 2.0-38.9 mg/kg, respectively). Based on the results of a regression analysis, silicate clays, carbonates and, especially, Fe oxides contribute positively to Olsen P 0.01 and Olsen P 0.03 . The fact that, as shown here, Olsen P at a specific solution P concentration can be predicted from easily measured soil properties can facilitate decision-making in managing fertilizer P.

No abstract