Studies characterizing the effect of a century of plant breeding on physiological traits are highly needed to identify candidate traits for future improvement in maize (Zea mays L.). A global evaluation of kernel weight progress over time requires the assembly of large and reliable data documenting genetic improvements in this trait across commercial breeding programs in different regions. We compiled a global dataset of yield and yield components from 34 published and unpublished studies comparing two or more maize cultivars from different decades of commercial release under field conditions. The dataset includes 750 entries of kernel weight data (requirement to be included in the systematic review), of which 642 and 666 include data entries of grain yield and kernel number, respectively. We also extracted the metadata describing experimental site information, agronomic management practices, and genotypic information. This dataset can be useful to identify trends of yield improvement across management conditions, with proper consideration of the trade-off between kernel number and kernel weight in maize.

This file contains data on kernel weight, kernel number, and grain yield for maize cultivars from different decades of commercial release collected from published literature. Missing values or information of studies are represented by "n/a" in the data. References of included studies in the data are described in the associated README file.