Ferruginous nodules sorb significant amounts of available soil and fer
tiliser phosphate. The effect of this sorption on phosphorus availabil
ity of an agricultural soil was tested by sequential extraction and by
exhaustive cropping with millet (Pennisetum typhoides) in a greenhous
e trial following fertilisation of the original soil containing 70% no
dules and of prepared samples containing various mixes of separated so
il fines and nodules. Phosphorus sorption maxima by the soil fines and
nodules were 190 mg kg-1 and 380 mg kg-1 respectively. Samples of fin
es and nodules which had sorbed 110 and 194 mg kg-1 were submitted to
8 successive extractions with 0.01 M KCl, after which P desorption amo
unted to 117 mg kg-1 and 103 mg kg-1 respectively. Hysteresis between
sorption and desorption was negligible for the soil fines and increase
d with increasing nodule content of the samples. In the greenhouse exp
eriment, P uptake at the first cropping was highest in the soil fines
at all levels of phosphorus applied. Subsequent croppings, however, sh
owed higher P uptake in the concretionary soils. These results indicat
e a higher initial P release from the soil fines with cropping followe
d by an earlier exhaustion of phosphorus. At the end of the greenhouse
experiment, yields were low in spite of the large quantities of P sti
ll remaining in the soils. Phosphorus fractionation showed that, of th
e P left in the soil after cropping 20% was in labile, 29% in Fe or Al
-associated, and 51% in low-availability forms.