This inquiry tested the hypothesis (Byrne, 1996) that young children who do not yet comprehend that the elements of alphabetic writing represent phonemes link writing to language at the level of morphemes. This study asked US preschoolers to write words that varied in the number of morphemes and the number of syllables that they included. It distinguished a group of 50 children who employed letters that represented phonemes in the intended words no more often than expected by chance (mean age = 4 years, 9 months). These prephonological spellers did not generate longer spellings for two-morpheme words like teacup than for one-morpheme words like napkin, albeit the length of their spellings was impacted by the number of letters that they employed to spell the formerly presented word and by the order of the word in the experiment. The results propose that the length of prephonological spellers’ productions is not affected by the linguistic length of a message in phonemes, syllables, or morphemes, and they do not support the idea that these children display a special sensitivity to morphemes.