Marbled Godwits (Limosa fedoa) do not present obvious sexual dimorphism, although females are known to be slightly larger than males, on average. During the nonbreeding seasons between August 2007 and April 2009, a noninvasive technique was tested that might permit phenotypic identification of sex in a large sample of individuals from the most important wintering area, Guerrero Negro, México. Blood samples (n = 577) were used for molecular sexing. For phenotypic sexing, culmen length was selected because it had the greatest sexual dimorphism. A normal distribution curve for culmen length of each sex and the overlapping area between the two curves was calculated and used to minimize the probability of error in identifying the sex of each individual. Using the criterion that males had a culmen length ≤ 106 mm and females ≥ 108 mm, the sex of 96.4% of the individuals would be correctly identified, 2.8% would be regarded as indeterminate and 0.8% would be sexed erroneously. A discriminant analysis using bill length and other commonly used phenotypic measurements did not improve success in predicting sex. The results indicate that culmen length is a useful measure for sex determination in Marbled Godwit. Identification of the sex of Marbled Godwits in the field offers new possibilities to study life histories of both sexes. In Guerrero Negro, the sex ratio of Marbled Godwits was heavily male-biased biased (2:1), suggesting latitudinal segregation by sex.