In order to find out the possibility of simplification of weeding operation on upland rice field this experiment was carried out in the following experimental design: the experimental main factor was 4 levels of the crop stage, when the competition against weed was comme ncedafter weed-free period, that were, the sowing stage (0L), the fourth leaf stage (4L), the six th (6L) and the eighth (81), and subplot factor was 2 levels of distance between the row, that were, 60cm. (hand sowing: A) and 30cm. (machine sowing: B). In addition to the levels mentioned aboved, we have such two other levels as plot completely weeded away (R) and plot only constituted of weeds (W). These were treated in Split-Plot Design with 3 replication. The results obtained were as follows. (1). Weeds (mainly Amaranthus retreflexus L.) were controlled more effectively by upland rice itself after weed-free period as the weed-free period on the field was continued longer. Also this tendency was more remarkable in narrow row distance (B) than in wide (A). The effect of row distance on weed control become more clear about the initiation srage of the internode elongation of upland rice. And it seems to be most distinguishable on 4L and 6L treatments. It is evident that this may have been induced from the light competition between weeds and uplan rice, and caucauced to the difference of "Light Intercepting Capacity" of upland rice among each treatment. (2). The crop yield in each of treatments may be arranged in such a order as B>A or 8L>R>6L>4L>0L. (3). Considering from these results, it may be conclucluded that upland rice should be sown in 30cm. row destance rather than 60cm. and raised cleanly in weed-free condition up to 4th or 6th leaf stage. It was discussed how to make the weeding pracice on the mechanized crop production simple and effective, in which we have ever had to waste the large amount of our labour and money.