This paper documents an externality from the agricultural use of the most popular herbicide in the world -- glyphosate -- on birth outcomes of surrounding populations. Our empirical strategy explores the initial regulation of genetically modified seeds in Brazil, the potential gains in productivity from adoption of genetically modified soybean, and the fact that glyphosate was strongly complementary to the first generation of genetically modified soybean seeds introduced in the country. We focus on the identification of the subclinical effects through contamination of water for populations distant from the original locations of use. Our strategy relies heavily on the direction of water flow within water basins to achieve identification. We detect a statistically significant deterioration in birth outcomes for populations receiving water from locations that expanded glyphosate use. Despite ongoing controversy, little is known about the externalities of herbicides in general on human populations at large. We provide evidence on this type of externality for the case of glyphosate contamination through water.