Experimental evidence for limitation of net primary productivity (NPP) by nitrogen (N) or phosphorus (P) in lowland tropical forests is rare, and the results from the few existing studies have been inconclusive. To directly test if N or P limit NPP in a lowland tropical wet forest in Costa Rica, we conducted a full factorial fertilization experiment (4 treatments × 6 replicates in 30 × 30 m plots). We focused on the influence of tree size and taxa on nutrient limitation, because in these forests a wide variety of tree functional traits related to nutrient acquisition and use are likely to regulate biogeochemical processes. After 2.7 years, a higher percentage of trees per plot increased basal area (BA) with P additions (66.45% ± 3.28% without P vs. 76.88% ± 3.28% with P), but there were no other community-level responses to N or P additions on BA increase, litterfall productivity, or root growth. Phosphorus additions resulted in doubled stem growth rates in small trees (5–10 cm diameter at breast height (dbh); [P ≤ 0.01]) but had no effect on intermediate (10–30 cm dbh) or large trees (>30 cm dbh). Phosphorus additions also increased the percentage of seedling survival from 59% to 78% (P < 0.01), as well as the percentage of seedlings that grew (P = 0.03), and increased leaf number (P = 0.02). Trees from Pentaclethra macroloba, the most abundant species, did not increase growth rates with fertilization (P = 0.40). In contrast, the most abundant palms (Socratea exorrhiza) had more than two times higher stem growth rates with P additions (P = 0.01). Our experiment reiterates that P availability is a significant driver of plant processes in these systems, but highlights the importance of considering different aspects of the plant community when making predictions concerning nutrient limitation. We postulate that in diverse, lowland tropical forests “heterogeneous nutrient limitation” occurs, not only driven by variability in nutrient responses among taxa, but also among size classes and potential functional groups. Heterogeneous responses to nutrient additions could lead to changes in forest structure or even diversity in the long term, affecting rates of NPP and thus carbon cycling.