The English language has 11 Basic Color Terms (BCTs), including a word for BLUE, but this is not true of many other languages. We have argued that worldwide variation in the use of words for BLUE is partly due to a type-III (blue-yellow) color vision deficiency caused by phototoxic damage to the eye by UV sunlight. Key evidence comes from the strong relation between UV exposure and words for BLUE (see: Lindsey & Brown, Psychol. Sci. 2002). However, our critics have argued that the languages we reviewed, being from published studies of language and color, could be biased towards languages with “interesting” BCTs. Here, we report a new study of the geographical distribution of words for BLUE. METHODS: We examined 280 languages listed on the website www.yourdictionary.com. They were not chosen by us, and they were not chosen for their “interesting” BCTs. Of the 280 listings, we identified 220 distinct languages that were spoken in 1500 c.e. and are still spoken today. We obtained color names in 202 (92%) of them by looking up translations of the 11 BCTs in printed or online dictionaries or by consulting a scholar or native speaker. We determined the UV dose from the latitude/longitude where the language is spoken using UV-B data from NASA. RESULTS: The prevalence of BLUE was highly related to the average daily dose of UV from sunlight: below 3350 J/m2/day (e.g. Istanbul), 90% of languages in our sample had distinct, native terms for BLUE. At higher doses, the prevalence of BLUE fell off approximately linearly. Only 11% of languages spoken at the highest UV doses (>5400 J/m2/day; e.g. Sri Lanka) included terms for BLUE. The frequency of BLACK or GREEN to mean BLUE increased with increasing UB dosage. CONCLUSIONS: The likelihood that a language will have a distinct native word for BLUE is inversely related to the average daily dose of UV from sunlight at the locality where it is spoken. This result is specifically predicted by our phototoxicity hypothesis.