Italian sixth graders, with and without dyslexia, read pseudowords and low‐frequency words that include high‐frequency morphemes better than stimuli not including any morpheme. The present study assessed whether morphemes affect (1) younger children, with and without dyslexia; (2) spelling as well as reading; and (3) words with low‐frequency morphemes. Two groups of third graders (16 children with dyslexia and dysorthography and 16 age‐matched typically developing children) read aloud and spelt to dictation pseudowords and words. Pseudowords included (1) root + suffix in not existing combinations (e.g. lampadista, formed by lampad‐, ‘lamp’, and ‐ista, ‘‐ist’) and (2) orthographic sequences not corresponding to any Italian root or suffix (e.g. livonosto). Words had low frequency and included: (1) root + suffix, both of high frequency (e.g. bestiale, ‘beastly’); (2) root + suffix, both of low frequency (e.g. asprigno, ‘rather sour’); and (3) simple words (e.g. insulso, ‘vapid’). Children with dyslexia and dysorthography were less accurate than typically developing children. Root + suffix pseudowords were read and spelt more accurately than non‐morphological pseudowords by both groups. Morphologically complex (root + suffix) words were read and spelt better than simple words. However, task interacted with morphology: reading was not facilitated by low‐frequency morphemes. We conclude that children acquiring a transparent orthography exploit morpheme‐based reading and spelling to face difficulties in processing long unfamiliar stimuli. Copyright © 2017 John Wiley & Sons, Ltd.