Two field experiments were conducted at the Corozal Agricultural Experiment Station, Puerto Rico, in 2000 and 2001 to evaluate four herbicide treatments in direct-seeded and transplanted calabaza [Cucurbita moschata Duchesne (Poir)] cv. Soler. Herbicide treatments were 1) glyphosate (1.75 U ha) preplanting (PP); 2) glyphosate (1.75 L/ha, PP) followed by sethoxydim (0.31 kg ai/ha) postemergence from five to nine weeks after planting; 3) clomazone (0.50 kg ai/ha, PP); 4) mixture of clomazone (0.50 kg ai/ha) + glyphosate (1.75 L/ha, PP); and 5) control with hand-weeded cultivation. Predominant weeds in the experimental area were large crabgrass [Digitaria sanguinalis (L.) Scop.], jungle rice [Echinochloa colona (L.) Link], wild poinsettia (Euphorbia heterophylla L.), and niruri (Phyllanthus niruriL.). At three and six weeks after planting, grasses and total weed density were usually higher in 2000 than in 2001. In 2000, preplant clomazone alone or in mixture with glyphosate was more effective than glyphosate alone in reducing grasses and total weed densities at three weeks after planting. Differences in weed densities in general were non-significant among herbicide treatments in 2001. Also in 2001 hand-hoe cultivation was more effective reducing grasses and total weeds than preplant mixture of glyphosate plus clomazone or preplant glyphosate alone. Differences in weed dry weight at the end of the season as well as in calabaza yields were non-significant for herbicide treatments. Transplanted calabaza produced higher mean yields than direct-seeded calabaza both years. This result indicates that transplanted calabaza may compete better than direct-seeded calabaza under low weed population levels (two to 23 plants per 0.5 m2), given the advantage of the early establishment of transplants before appearance of weeds.