Bottles were cleaned as described above, and stored in 0.01 M Q-HCl. Prior to use, they were thoroughly washed with UHP water and rinsed with the seawater sample. Samples were collected from 10 l modified Niskin bottles (General Oceanics) designed for trace metal work mounted on an epoxy-coated stainless steel rosette frame. Stringent clean sample handling techniques were followed. Unfiltered samples were collected in HDPE bottles and acidi®ed to pH 2 with 250 ml of Q-HCl (9.0 M) per 250 ml of sample. The Fe(III) reducing reagent was added to acidi®ed samples (final sulphite concentration 100 mM) and allowed to react for at least 4 h. The samples were analysed and calibration performed by standard additions.
Three 10 ml aliquots of reduced sample were transferred to clean 25 ml polystyrene screw-capped vials (Sterilin, Merck BDH). Sub-sample 1 was left alone, whilst sub-samples 2 and 3 were spiked with a small volume (20±50 ml) of Fe(II) standard (variable concentration) to achieve standard additions ranging from 0.2±5.0 nM. The sub-samples were mixed and analysed in order of increasing concentration.