Domestic fowl in small groups are assumed to establish hierarchical systems based on individual recognition. Conversely, interactions in large groups are modulated by badges of status. Previous studies suggested that birds differing in phenotypic appearance (PA) became targets for aggression, possibly mistaking altered PA for badges of status. We evaluated the impact of altering PA on 0, 30, 50, 70 or 100% of the birds’ house at three experimental group sizes (GS). Tested GS were 10, 20 or 40 (8 birds/m2, 3 pens/GSxPA, 45 total). Thus, for each GS we had groups initially homogenous (100U, U = Unmarked; 100M, M = Marked), or heterogeneous M and U phenotypes coexisting in different proportions: 30M/70U, 50M/50U, and 70M/30U, remaining unchanged until 33 weeks of age. Then, homogeneous groups (100U and 100M) were sequentially altered by marking or unmarking 30, 50 and 70% of birds at 34, 38 and 44 weeks, respectively. Aggressive interactions were observed before applying changes at 27–28 weeks (T0), and after each sequential PA change on week 35–36 (T1), 39–40 (T2) and 45–46 (T3). Frequency of aggressive interactions in altered groups at T1, T2, and T3 were compared with non-altered heterogeneous controls. Results indicate no differences across initial PA and GS treatments (T0; P>0.05). However, aggression escalation was observed at T1 immediately after the first PA manipulation (Tukey P<0.05 altered vs controls). Aggression in altered groups remained high at T2 when compared to controls (Tukey, P<0.05), although by T3 interactions declined to almost initial levels (Tukey, P>0.05 altered vs controls). Aggressive interactions at T1 and T2 were predominantly directed from un-altered towards recently altered birds, irrespectively of their initial phenotype and of the GS. These results demonstrate that a sudden change in PA affects group dynamics. Altered birds were exposed to escalated aggression even in small groups, where individual recognition was presumed.