The present study investigates Ehri's (Ehri & Wilce 1985; Scott & Ehri 1990) hypothesis that knowledge of the alphabet enables children to learn to read by processing and storing letter-sound relations in words. In particular, it examines whether letter-name knowledge facilitates the learning of spellings in which the names of one or more letters can be heard in the pronunciation of the words. Preschool children who could not read any word out of context were divided into two groups on the basis of their ability to name the letters of the alphabet: one group knew the names of the letters while the other did not. Both groups were taught to read two types of simplified spellings: visual spellings, that is, spellings whose letters did not correspond to sounds in the pronunciations of the words but which were visually more salient (e.g., XQKO for the word ‘cerveja’), and phonetic spellings, that is, spellings whose letters corresponded to sounds in the pronunciation of the words (e.g., CRVA for the word ‘cerveja’). In all phonetic spellings, the name of at least one letter could be clearly heard in the pronunciation of the words. Results corroborated Ehri's hypothesis. The children who did not know the names of the letters learned to read the visual spellings more easily than the phonetic ones. On the other hand, the children who knew the names of the letters showed the opposite pattern, that is, they learned the phonetic spellings more easily than the visual ones.