Technical Abstract: With rising energy costs, deep tillage carries an increasing cost to many Coastal Plain management systems; therefore, deep tillage needs to be reevaluated in terms of frequency and implement type. In 2002 and 2003, 3 implements in 4 tillage treatments were evaluated for effectiveness in increasing cotton (Gossypium hirsutum L.) yield at two sites when non-inversion deep tillage was performed annually or not. The 4 treatments used a straight-legged subsoil shank with bedding, a straight-legged subsoil shank and strip tillage without bedding, a Paratill and strip tillage without bedding, and a Terra-Max and strip tillage without bedding. In the second year of the experiment, treatments were split; half of each plot was tilled again and half was not. Treatments were compared to a control that was not tilled. Tillage generally decreased soil penetration resistance and improved yield but differences were significant only about half of the time. No single tillage treatment significantly reduced penetration resistances better than the others at both sites or for both years. Treatments that were tilled the first and not the second year did not significantly reduce penetration resistance probably because of lack of re-compaction based on an exceptionally dry first growing season. Tilling the second year improved yield but only marginally. Producers would do well to decide whether or not to till after a dry year (and possibly subsequent years) on a case-by-case basis, basing their decision on re-compaction.