SUMMARY 1. We examined microbial respiration among streams in lowland Costa Rica comprising a natural phosphorus gradient (5–350 μg SRP L−1) resulting from variable inputs of solute-rich (e.g. P, SO4 and Cl) groundwater.
2. Microbial respiration rates were determined by measuring oxygen change in situ in nine low-order streams on three substrate types: mixed leaves collected from the stream bottom, conditioned Ficus leaves and sediments.
3. Respiration rates on both leaf types were positively related to phosphorus and negatively related to N : P ratios. Microbial respiration rates on sediments were not related to any of the variables [i.e. soluble reactive phosphorus (SRP), N-NO3 and N : P] measured.
4. Respiration rates on newly colonised Ficus leaves formed an asymptotic curve increasing to a plateau, suggesting that saturation with phosphorus occurred at concentrations <15 μg SRP L−1.
5. To test the hypothesis that phosphorus was the main solute in solute-rich water that was driving observed differences in microbial respiration rates, we artificially enriched a small stream with phosphorus and measured changes in respiration before and after enrichment.
6. Experimental phosphorus enrichment produced increases in respiration rates similar in magnitude to those observed in the nine streams forming the natural phosphorus gradient, supporting our hypothesis that phosphorus was the major variable driving interstream differences in microbial respiration rates. Respiration rates were higher in this study than those reported for most other tropical streams and rivers with the exception of those reported for tropical Asian streams.
7. Results indicate that variations in phosphorus concentrations can potentially affect patterns of microbial respiration rates at a landscape level via differential inputs of solute-rich groundwater into streams.