Primocanes of `Marion' trailing blackberry plants (Rubus spp.) were suppressed by cutting them off at ground level in either late April, May, June, or July 1991 and 1992. A control was included in which primocanes were not cut. Four canes per plant were trained in either August or February, with all other canes being removed and measured. Yield data were collected in 1992 and 1993, after which yield components were measured. Cane diameter was greatest for unsuppressed plants and declined with later primocane removal date. Cane length was greatest for unsuppressed and April-suppressed plants. Internode length decreased and main cane percent budbreak increased with later suppression date. Cane number and total main cane length per plant were increased in April-, May-, and June-suppressed plants in 1992 and for April- and June-suppressed plants in 1993. Consequently, yield of April-suppressed plants exceeded that of unsuppressed plants in 1992. Yield of April-, May-, and June-suppressed plants exceeded that of unsuppressed plants in 1993. August-trained plants yielded 46% more than February-trained plants, primarily because of higher percent budbreak on main canes. August-trained plants also produced longer canes with more nodes and a greater number of fruit per main cane lateral.