Most children learn a language such as Mandarin Chinese that uses lexical tone to communicate meaning. This study aimed to examine the phonological specificity of tone representations in monolingual and bilingual learners of Mandarin. Two age-groups were tested: toddlers (2.5 to 3.5 years) and preschoolers (4 to 5 years; N = 80). Using a preferential looking paradigm, children were presented with correct and mispronounced (Cantonese or Mandarin tone substitutions) forms of familiar words. In the aggregate, participants demonstrated preferential fixation to visual targets (naming effects) for correctly pronounced words, words mispronounced with native (Mandarin) tone substitutions, and for words mispronounced with nonnative (Cantonese) tone substitutions. However, naming effects were greater for correct pronunciations and for Cantonese tone mispronunciations relative to Mandarin tone mispronunciations. There were no interactions of responses to each trial type with age or with language exposure. Results are discussed in terms of possible determinants of tone sensitivity in early childhood.