This dissertation used letter detection and masked priming to address four questions: Are graphemes or letters more fundamental in low-level reading processes? How does alphabetic-processing knowledge manifest in different languages? Do bilinguals transfer such knowledge across languages? And do young children also show such effects?Some researchers have recently revived an old hypothesis in which graphemes, not letters, are the fundamental, perceptual reading unit. This can be tested by looking at congruency effects in a letter-detection task with masked priming. Six groups participated: Spanish and English monolingual adults; Spanish- and English-dominant bilingual adults; Spanish-dominant bilingual children; and English monolingual children. The experiments with adult monolinguals tested the letter- against the grapheme-as-percept hypothesis. The experiments with developing bilinguals examined whether they would transfer alphabetic-processing knowledge from L1 to L2. And the experiments with English monolingual children probed how early congruency effects with masked primes might occur.Participants responded YES or NO depending on the presence of letters in targets. Both congruent and incongruent masked primes preceded the targets. Among the congruent primes, some contained double vowels, and others single vowels. Assuming letters are fundamental, single- and double-vowel primes in both languages should facilitate and inhibit reactions equally. Assuming graphemes are fundamental, single-vowel primes in English, but not Spanish, should facilitate and inhibit more because double vowels are digraphs in English, and should therefore conceal the identity of their component letters. Bilinguals should show L1-like effects in L2 if they transfer alphabet-specific processing knowledge. Young children should simply show congruency effects if they are able to process letter information automatically.The results with Spanish and English monolinguals suggested that graphemes do exert an effect on the task, but only after letters are perceived. This has major implications for models of proficient reading. The results also suggest that Spanish readers do not construct graphemes from letters, but rather syllables and abstract syllable structure. Bilinguals showed evidence of L1-L2 transfer at low levels of L2 proficiency. This has implications for transitional bilingual education programs. And young children showed congruency effects, which provides another link in establishing the connection between literacy development and proficient reading.