Two experiments examined Czech children's comprehension of grammatical number marking in verbs. Children were presented with picture pairs involving one or multiple participants in the same action, and were asked to point to the picture described by a recorded sentence. Experiment 1 (N = 72, age 3;0-4;7) tested four types of sentences, some of which marked the number of participants on the verb only. There was a marked increase in comprehension scores with age and receptive vocabulary scores in singular, but not in plural sentences. This pattern was present regardless of how number was marked. Experiment 2 (N = 45, age 3;4-4;9) replicated the key findings of Experiment 1. The results suggest that comprehension of morphological number markers remains difficult even when children routinely produce such markers. However, when compared with results from other languages and experimental paradigms, it appears that the effect is due mainly to children's comprehension of the task and not limitations in their grammatical knowledge.