A 17 acre (6.9 ha) agroforestry research and extension alley cropping trial was established at the Center for Environmental Farming Systems in Goldsboro, North Carolina in January 2007, with a randomized block design with five replications. Loblolly pine (Pinus taeda), longleaf pine (Pinus palustris) and cherrybark oak (Quercus pagoda) were planted in staggered rows, with each species planted for 140 ft (43 m) per replication. Crop land alleys of 40 ft or 80 ft (12.2-24.4 m) wide were left between the tree rows. Crops of soybeans (Glycine max) and corn (Zea mays) were planted in alternating years since establishment. As of 2011, survival rates were 93% for cherrybark oak, 88% for longleaf pine and 97% for loblolly pine. Average tree diameter at ground level was 1.0 in (2.5 cm) for cherrybark oak, 2.1 in (5.3 cm) for longleaf and 3.2 in (8.1 cm) for loblolly. Heights averaged 4.6 ft (1.4 m) for cherrybark oak, 5.2 ft (1.6 m) for longleaf pine and 10.4 ft (3.2 m) for loblolly pine. Growth, yield and economic projections for traditional timber production indicated that species volumes and values tracked the height and diameter relationships measured on the site. Loblolly pine had the largest projected internal rate of return, at 7.2%, followed by longleaf pine with pine straw harvests at 5.5%, longleaf without pine straw at 3.5% and cherrybark oak at 1.9%. There might be more loss in crop and silvopasture production with loblolly, however, and production of pine straw for longleaf or game mast for cherrybark oak may offer other benefits. Crop yields on the sandy soils on the site were very poor during the 4 years observed, which had a series of droughts and floods. These led to net financial losses in those years for the demonstration site, but state-wide average farm budget returns did show moderate profits. The results support the merits of agroforestry systems in the upper South to diversify income and reduce financial risks.