In the reclamation of scrublands of Galicia in NW Spain, fertilization is necessary to establish pastures. The area is particularly responsive to P fertilizer. In order to maximize the efficiency of use of fertilizer in planning fertilization programs, it is necessary to take into consideration the residual as well as the freshly added P.

Phosphate fertilization experiments were conducted in four sites in Galicia having a range in soil properties. The treatment design allowed estimation of residual P one year after fertilization in terms of freshly applied P equivalents.

Four non-linear models were fitted to the data from the experiments after it was determined that residual soil P one year after fertilization was proportional to the amount of P applied. The estimated proportion of P remaining one year after application appeared to be independent of the model chosen. The Mitscherlich performed best from the standpoint of average residual mean squares, total rank, residual behavior after the fit and ease of fit.

The percent of fresh fertilizer equivalent estimated one year after addition ranged from 60–90 for soils formed from granite, 30–60 for soils from slates, and 30 for soils over schist and gabbros.