In any mechanized operation is the common occurrence of losses, and in the case of the harvestis important that these be quantified in order to minimize them. The work aimed was evaluated theinfluence of spatial quantitative losses (soil, plants and total) in operation mechanically harvestedcotton, using frames with collection areas 1.125 m², 2.25 m² and 4.50 m². Samples were collected inthe municipality of Ipameri, GO, on a plot of 1.4 ha, being demarcated 41 points, spaced at 50 mx 9m. The crop was harvested by a combine harvester John Deere, model 9960, with five lines platform.The samples were collected by placing the frame on the ground, collecting all cotton lying onthe ground and then made up the collection that remained in the cotton plant after the passageof the harvester. Data were analyzed using descriptive statistics and geostatistics for making themaps. Losses had found large spatial variability, with high indices considered for the mechanicalharvesting of cotton, predominantly losses on losses in plants in soil. The frame of 4.50 m² showed thesmallest variance and coefficient of variation than the frames of 1.125 and 2.25 m².Em qualquer operação mecanizada é comum a ocorrência de perdas, e no caso da colheita o importante é que essas sejam quantificadas de forma a minimizá-las. Objetivou-se avaliar espacialmente as perdas quantitativas (no solo, nas plantas e totais) na operação de colheita mecanizada de algodão, utilizando-se armações de coleta com áreas de 1,13 m², 2,25 m² e 4,50 m². Foram coletadas as amostras no município de Ipameri, GO, em um talhão de 1,4 ha, sendo georreferenciados 41 pontos, distanciados em 50 x 9 m. A colheita foi realizada por uma colhedora John Deere, modelo 9960, com plataforma de cinco linhas. As amostras foram coletadas colocando-se a armação sobre o solo, coletando o algodão caído no solo, e o remanescente na planta após a passagem da colhedora. Os dados foram analisados pela estatística descritiva e pela geoestatística para confecção dos mapas de variabilidade espacial. As perdas encontradas apresentaram grande variabilidade espacial, sendo os índices considerados altos para a colheita mecanizada do algodão, predominando as perdas nas plantas sobre as perdas no solo. A armação de 4,50 m² apresentou a menor variabilidade do que as armações de 1,13 e 2,25 m². Spatial variability process of the losses in the mechanized harvesting of cotton withdifferent frames collectionAbstractIn any mechanized operation is the common occurrence of losses, and in the case of the harvestis important that these be quantified in order to minimize them. The work aimed was evaluated theinfluence of spatial quantitative losses (soil, plants and total) in operation mechanically harvestedcotton, using frames with collection areas 1.125 m², 2.25 m² and 4.50 m². Samples were collected inthe municipality of Ipameri, GO, on a plot of 1.4 ha, being demarcated 41 points, spaced at 50 mx 9m. The crop was harvested by a combine harvester John Deere, model 9960, with five lines platform.The samples were collected by placing the frame on the ground, collecting all cotton lying onthe ground and then made up the collection that remained in the cotton plant after the passageof the harvester. Data were analyzed using descriptive statistics and geostatistics for making themaps. Losses had found large spatial variability, with high indices considered for the mechanicalharvesting of cotton, predominantly losses on losses in plants in soil. The frame of 4.50 m² showed thesmallest variance and coefficient of variation than the frames of 1.125 and 2.25 m².