Agricultural Sciences: Techniques and Innovations Vol. 12

Heterotic grouping is central to maize hybrid breeding because it structures parental pools, guides tester choice and concentrates crossing effort on combinations with a higher probability of superior performance. Yet the term is frequently used imprecisely, and molecular clusters, pedigree classes and combining-ability groups are often treated as interchangeable even though they capture different biological and operational signals. This critical narrative review evaluates the conceptual basis, regional development, classification methods and genomic future of heterotic grouping in maize. Literature published principally from 1990 to 31 May 2026 was selected through live searches of accessible scholarly indexes and DOI metadata sources, supplemented by citation chaining and verification against official article records. The evidence indicates that a heterotic group is best understood as a breeding population defined by repeatable cross-performance relationships, not as a taxonomic unit or a neutral-marker cluster. Long-established temperate patterns, including Stiff Stalk versus non-Stiff Stalk in North American dent maize and Dent versus Flint in Europe, show comparatively strong alignment among pedigree, genomic differentiation and hybrid performance because recurrent selection and germplasm recycling have reinforced those divisions. In tropical, subtropical and recently assembled breeding material, agreement is weaker: genetic distance can describe ancestry and avoid close relatives, but it does not consistently predict heterosis or specific combining ability. Testcross evidence remains the most direct basis for operational assignment, although results depend on tester choice, target environment, maturity, stress regime and trial precision. Heterotic group-specific and general combining ability methods can improve classification in some populations, while genomic prediction generally offers greater practical value than simple marker distance when adequate training relationships and multi-environment data are available. The most defensible strategy is therefore integrative and dynamic: combine pedigree and genomic structure with replicated testcross performance, update assignments as programmes evolve, and manage within-group diversity as deliberately as between-group divergence.