The mammalian neocortex is an exquisite, highly organized brain structure composed of hundreds of subpopulations of neurons and glia, precisely connected to enable motor control, sensory perception, information integration, and planning. Unique molecular, structural, and anatomical neuronal properties underlie diverse functionality, endowing much of the neocortex’s complex processing power. Neocortical size correlates with information processing capacity, suggesting that increased neuronal number and diversity begets increased sophistication. One excitatory projection neuron type, callosal projection neurons (CPN), has disproportionately expanded with cortical size increase. CPN directly connect homotypic regions of the two neocortical hemispheres by sending axons via the largest white matter fiber tract in the brain, the corpus callosum (CC), allowing quick relay, integration, and comparison of information. In humans, the CC contains over 300,000 axons, CPN have been centrally implicated in autism spectrum disorders, and absence or surgical disruption of CPN connectivity in humans is associated with defects in abstract reasoning, problem solving, and generalization. Therefore, CPN are critical to complex brain functions, and their diversity likely contributes to these roles. Work presented in this dissertation addresses molecular controls over CPN development, specifically genes that are expressed by, and function in, particular subpopulations of CPN. While much progress has been made in identifying molecular controls over neocortical arealization, lamination, and broad subtype specification, CPN diversity has remained largely unaddressed. Therefore, this work begins by identifying genes more highly expressed in CPN than other closely related projection neuron populations, and uncovers molecular diversity within CPN. From this molecular diversity, functional analysis of three candidate molecular controls over CPN subtype diversity follows. Cited2 acts broadly in neocortical progenitor development and postnatally in refining somatosensory CPN identity. Caveolin1 identifies a population of CPN with dual axonal projections. Tmtc4 is mutated in human CC disease and can function in CPN axonal development. These analyses of CPN molecular diversity in mouse then expand to an investigation of which molecular subpopulations are conserved, expanded, or uncommon between rodent and primate, allowing both for comparative evolutionary theories of CPN function, and indicating which CPN populations critical for human brain function can be best studied in rodent models.