The aim of the present research was to reexamine the production of reflexive and anti-causative se-verbs in Serbian at different stages of language acquisition after a nine-month period from the first experiment (Ilić, 2019). The results of the first experiment suggest the following order of acquisition: lexical reflexive > true reflexive > anti-causative verbs. Moreover, they speak in favour of the Continuity Hypothesis (Snyder et al., 1995; Snyder & Hyams, 2015), since the children were able to produce reflexive forms correctly from the earliest tested age. The verbs which were tested in both experiments were true reflexive (e.g. kupati se ‘wash oneself’), lexical reflexive (e.g. vrteti se ‘spin’) and anti-causative se-verbs (e.g. otvoriti se ‘open’). A total of sixty subjects belonging to three age groups (40-51, 52-63, 65-77 months, twenty participants in each group) took part in the research. The data collection technique was a verb elicitation task, and the children were asked to name the activities presented in the visual stimuli. The number of tested verbs was the same for each verb type (six per group). The production of verbs across age groups improved in comparison with the first experiment. The production of lexical reflexive verbs was significantly better than the production of anti-causative verbs in the youngest group. Therefore, the results of the follow-up experiment have confirmed the suggested order of acquisition. However, no differences were found in the two older groups, which indicates that the production of more complex verb types improves considerably between the ages of four and five.