Although grain yield (GY) in maize (Zea mays L.) is the product of both kernel number (KN) per unit area and kernel weight (KW), the latter has usually been considered the least variable component. However, sink strength enhancement by genetic improvement has meant that KW itself is more likely to limit yields in modern hybrids. This research studied the physiological mechanisms underlying KW by evaluating the effects of N availability on: 1) the sources of dry matter (DM) and N assimilates; 2) the determination of potential KW; and 3) the realization of final KW. To investigate how N availability affected source capacity, a 2-year field study (2016-2017) that combined N timing and N rate treatments was conducted. DM and N sources were calculated separately for the lag phase (R1-R3) and the linear phase (R3-R6). In the lag phase, substantial DM gains in leaves and stems occurred in both years, while N content gains were mostly detected in reproductive tissues. GY differences were consistently and mainly explained by KW variability, with reproductive tissues proving to be relatively stronger sinks for N than for DM during the lag phase. To understand how N availability affected both the determination of potential KW in the lag phase and the realization of final KW in the linear phase, three field experiments testing N rates, plant densities and N timing applications on a single hybrid were conducted (2017-2019). Average GY gains from zero N to the maximum N rate tested exceeded 8 Mg ha-1. Endosperm cell number (ECN), an indicator of potential KW, was determined at different timings (from 9 to 17 days after silking -DAS-), and plant component biomass plus N samples were taken at V12, R1 and R3. Low N treatments consistently reduced ECN at 9, 10, 13, and 17 DAS. Positive KW responses at R6 to N rates were always explained by higher ECN. Each year, ECN was highly correlated with ear N allocation rate during the lag phase. Potential KW was not associated with plant growth rate per kernel during the critical period, and a positive relationship (rather than a trade-off) was found between KW and KN. Additionally, since kernels accumulate both DM and N assimilates during the linear phase, kernels were sampled weekly over a 9-10-week interval from the early R3 stage. Linear plateau models were then fitted to the resulting kernel DM and N data to obtain characterizing parameters. Increases in N supply improved all four parameters: effective grain-filling rate, grain filling duration, kernel N accumulation rate (KNAR) and duration. KNAR was strongly correlated with both final KW (r=0.96) and with whole-plant N uptake by R3 (r=0.80). Overall, this dissertation provides new evidences for the distinct indirect and direct roles that N plays in the physiological mechanisms that determine final KW in maize at high potential GY.