Triple superphosphate (TSP), Arad phosphate rock (APR), and "Longlife" superphosphate (LL) were compared for their effectiveness as P fertilisers on a ryegrass/clover sward in New Zealand to which lime had been applied at 0, 1.25, 2.5, and 51 ha -1 . The P fertilisers were applied annually for five (LL) or six (TSP and APR) years at 24 kg P ha -1 yr -1 ; there was also a nil P control. The trial design was a factorial of four P treatments × four lime treatments. DM yield responses were initially negligible but increased to a maximum of 27% by Year 5. Mean response to P fertiliser was 5% in spring and 16% in summer and autumn. There were significant DM responses to TSP and LL from Year 3 onwards and to APR from Year 4 onwards. Over the final three years, TSP and LL yielded significantly more DM than APR, but there was no significant difference in DM yield between TSP and LL. Herbage %P, herbage P uptake, and soil Olsen P were much more sensitive discriminators between P fertilisers than was DM production and showed significant differences between all fertilisers throughout most of the experiment, with the ranking of effectiveness being TSP>LL>APR>nil P. All measurements showed APR to be a relatively ineffective P fertiliser, giving only 41% of the response to TSP in Year 6 when averaged over all lime treatments. Herbage chemical analysis showed effectiveness of APR to be markedly reduced by 2.5 and 51 ha -1 lime. Calculations based on residual *Dr A. G. Sinclair died on 2 December 1996 while this paper was in preparation.
A97007 Received 3 February 1997; accepted 10 October 1997PR in soil at the end of the experiment indicated that dissolution rates of APR with nil, 1.25, 2.5, and 5 t ha -1 lime were 23%, 20%, 7%, and 9% per year, respectively. The slow dissolution of APR was attributed to its relatively low reactivity as measured by solubility in formic acid. The PR content of LL (North Carolina PR) also proved very ineffective, with an average dissolution rate of 13% per year which was unaffected by lime application rate. It is suggested that granulation of LL depressed dissolution of its PR component. Lime at 5 t ha -1 raised soil pH from 5.6 to 6.5 one year after application, with intermediate rates having a pro rata effect. Soil pH subsequently declined, but with considerable year-to-year fluctuations, at rates proportional to the lime application rate. There were no positive DM responses to lime but there were significant depressions in Years 5 and 6. Lime reduced herbage %P in APR treatments only. Lime significantly reduced Olsen P in control and all fertiliser treatments throughout the experiment, but only with APR was there an associated reduction in herbage %P.