A new dual channel flow injection analyser that can simultaneously analyse soluble reactive phosphorus and bromide in the field, has been used in an experiment to test the hypothesis that the phosphorus uptake length in Myrtle Creek, a small forested stream in the Australian Highlands, is influenced by the initial phosphorus concentration used in whole-stream release studies. The phosphorus uptake length was found to decrease with decreasing initial phosphorus concentration added; the uptake length was 98 m when an initial P concentration of 51.0-mu-g l-1 was used, 90 m with 21.7-mu-g l-1 and 63 m with 12.7-mu-g l-1. The estimated errors in the uptake lengths were 6-8%. Approximately 32% of the added phosphorus was retained in the 32 m study reach, with almost all (ca. 93%) of this retained phosphorus taken up by the sediments (microbial uptake plus physico-chemical adsorption) and only a small amount retained in transient storage zones.