There is increasing concern that due to land pressure and the need to maximize income, smallholder coffee farmers are increasingly being forced to cultivate in areas which are considered to be sub-optimal for coffee. Little is known about optimal coffee and tree combinations in these conditions and the degree to which crops and trees compete or are synergistic. In environmental conditions which were sub optimal for coffee cultivation in Nicaragua (1470 mm annual rainfall, 27 °C mean annual temperature and 455 m altitude compared to optima of 2000 mm, 23–24 °C and altitude between 1000 and 1400 m at that latitude, respectively), coffee and shade tree transpiration and soil evaporation were directly and separately measured in agroforestry (AFS) and full sun systems (FS). AFS was found to be a more efficient water user than FS because a greater proportion of rainfall was used by plant transpiration rather than being lost by soil evaporation. Plant transpiration accounted for 83% and 69% of evapotranspiration while soil evaporation represented 17% and 31%, in AFS and FS respectively. In AFS most of the water transpiration was due to coffee (72.5%) and much less by deciduous Tabebuia rosea (19%) and evergreen Simarouba glauca shade trees (8.5%). Furthermore, the study demonstrated the vastly different behaviour in water use by the shade trees. When in leaf, Tabebuia rosea transpired at four to six times the rate of evergreen Simarouba glauca, although crown sizes were similar. Contrasting precipitation between two consecutive years of study demonstrated that competition for water between coffee and shade tree occurred only in a severe dry season when coffee leaf water potential (LWP) reached its lowest values of -2.33 MPa in AFS. It was concluded that in most circumstances there was sufficient water for both coffee and trees, that coffee in AFS was a more efficient user of water than FS coffee, and that evergreen Simarouba glauca was more suitable as coffee shade tree co