Potassium is an essential nutrient required by plants in large quantities. It is present in soils in four pools, of varying availability to plants, which exist in dynamic equilibrium. K deficiencies have been observed in crops, such as cotton, maize and sorghum, in Vertisols in north-west of New South Wales, Australia. Application of K fertilizers in many of these soils exhibits no yield response. This study was undertaken in order to investigate the effect of soil depth and mineralogy on the response of maize and K dynamics following K fertilization. Vertisols were sampled from 2 depths (0-10 cm, 10-30 cm) from 4 sites in Northern NSW for use in a glasshouse trial. K treatments of 0, 100, 200 or 500 mg/kg K were applied in a randomized complete blocked design with 4 replications. Basal application of other major nutrients was applied to all pots. Three maize crops were grown in each pot for 7 weeks per crop. At each harvest shoot dry matter yield and K concentration were measured. Following the final harvest K present in soil exchangeable and non-exchangeable pools were measured. The original soil samples were analysed for general soil properties, mineralogical properties by X-ray diffraction, different forms of K and quantity intensity characteristics. Changes in different forms of K over the experiment were monitored and 4 distinct K dynamics mechanisms were defined. The results obtained showed definite stratification of K down the profile. There was no shoot yield response to applied K despite an increase in K uptake. Topsoils from sites with clay fractions dominated by interstratified smectite were more likely to fix K, while all subsoils had the capacity to fix K when high levels were applied. NaBPh4 extractable K was found to be highly correlated with available K and it is suggested that this test would be a more accurate method of measuring K than that currently employed.