Mycelial cord systems, of the basidiomycete Phanerochaete velutina, a common woodland saprotroph, were grown on unsterile soil in model laboratory microcosms from 4 cm3 wood inocula. Systems were supplied after 37 d with a fresh 4 cm3 beech wood ‘bait', placed behind the foraging colony margin. Systems were subject to dry shift (−0.056 MPa) or wet shift (−0.009 MPa) over an 11 d period either immediately after, or 20 d after baits were supplied. Controls were maintained at constant soil matric potential (−0.019 MPa). 57-d-old systems were supplied with NH4K2PO4 including 32P tracer within soil compartments local to inoculum or bait. Image analysis was used to quantify morphological responses to water regime and resource supply, and tracer movement monitored non-destructively with a scintillation probe for 57 d. 32P uptake was greatest when tracer was supplied local to the inoculum. Dry shift concurrent with bait supply caused system wide cord-thickening, prevented polarised growth towards the newly supplied bait, had a significant carbon (energy) cost compared to controls, significantly reduced 32P acquisition, and significantly increased 32P relocation to the bait. Wet shift concurrent with bait supply caused considerable loss of extra-resource mycelium in the unbaited region, resulting in highly polarised development along the bait-inoculum line, but did not affect 32P uptake and partitioning. Delayed wet shift caused swifter polarisation towards the bait, quantified in terms of fractal dimension, did not result in system wide regression of extra-resource mycelium, and resulted in correspondingly increased rates of 32P acquisition. Delayed dry shift prevented polarised growth towards the bait and had only transient effects on 32P uptake and partitioning. Results suggest that resource capture took priority over coordination of C reserves and reallocation of mycelial effort, and that mycelium colonising the new resource was more dependent on P translocate during desiccation stress.