Yellow-bellied Sapsuckers (Sphyrapicus varius) obtain phloem sap from clusters of holes that they peck in living trees. I examined trees that sapsuckers used for sap extraction in northern Michigan and tested several hypotheses to explain why they choose specific trees for attack and why they cluster their holes in one place on the bole of each of these focal trees. Sapsuckers preferentially attacked individuals of paper birch (Betula papyrifera), red maple (Acer rubrum), juneberry (Amelanchier sp.) and bigtooth aspen (Populus grandidentata). They made clusters of sap holes an average of 7.13 m from the ground and within 1 m of a live branch; most clusters were located above old holes or other wounds from previous years. Each new sap incision through the bark to the phloem stream was made above a previous one and was enlarged for an average of 3.1 days before being abandoned for a new, higher hole. The resulting long vertical chains of holes were made alongside others to form tight clusters of holes at a single spot on the tree. Sapsuckers did not select trees for sap extraction based on location relative to nesting sites or on microclimate conditions of water availability and tree density. Individual trees used for sap extraction did not have thinner bark, more moisture in bark samples, or larger crown, but they did score lower in an index of overall tree health. Experimental evidence suggested that sapsuckers cluster their holes to induce the accumulation of sap in bark that they will attack for future sap extraction. Thus, sapsuckers appear to overcome some of the difficulties in obtaining phloem sap by choosing specific species and individuals, clustering sap holes above previous wounds, and possibly by farming their resource througout the season by girdling the tree’s phloem stream with each successive sap incision. I attempted to duplicate the wounding techniques and patterns of wounding of sapsuckers but was unable to induce sap flow from the same or similar trees in the area.