The present paper focuses on the physiology of yield potential in wheat (Triticum aestivum L.), because breeding progress in yield potential has overtaken farm yield progress. The paper examines developments largely in the last 10 years seeking routes to higher yield potential. Lately this subject has come under pressure from two new imperatives: perceived slowing of genetic progress and ambitious functional genomics. Analysis of trials between 1996 and 2005 at the CIANO research centre in northwest Mexico suggests that yield potential progress in CIMMYT spring wheats has slowed to around 0·50% per year, but has not ceased there nor in winter wheats elsewhere. Meanwhile, in the last 10 years or so, physiological understanding has advanced somewhat. Increased kernel number/m2 remains strongly associated with genetic progress in grain yield, and new research reinforces the importance of spike dry weight (g/m2) at anthesis in its determination. Lengthening the spike growth period through manipulation of sensitivity to photoperiod looks promising, but more attention to kernels per unit of spike weight is also urged. With respect to plant height, an optimum somewhere between 0·7 and 1·0 m is accepted and we are moving away from infatuation with the Norin 10 dwarfing genes as a way of reaching that. What has not been achieved is good lodging resistance in all short spring wheats, nor a complete understanding of its physiological basis. New information is coming to light on the possible role of stored stem reserves at anthesis, for these reserves appear to have increased as yield potential has increased. Part of the benefit may be related to assimilate supply per kernel around anthesis, which new understanding suggests is important for maximum potential kernel weight. Nevertheless, results continue to suggest that despite more kernels/m2, the most recent wheats are still largely sink-limited during grain filling. Growing evidence from spring and winter wheat (and from rice and maize) now points to the importance of increased photosynthetic activity before and around flowering for recent genetic increases in yield potential. This opens up new possibilities for selection in field plots. Finally, attention is given to effects of weather on yield potential and recent advances in techniques for elucidating the physiological basis of genotype by year interactions. From physiological understanding such as described, traits can be suggested as possible selection criteria for yield potential. However, apart from the ACIAR/CIMMYT project looking at stomatal aperture-related traits as well as source and sink traits (Condon et al., in press; Reynolds et al., in press; van Ginkel et al., in press), there appear to have been few attempts to validate physiological (or morphological) selection criteria for wheat yield potential in the last decade, but recent promising results with spectral reflection indices could foreshadow more validation work. This contrasts with efforts to improve the performance of wheat (and maize) under water-limited conditions, and with the new plant type and super rice approaches of IRRI and China, respectively. Such research could be mapped out for wheat yield potential improvement, and could lead to more efficient breeding for yield potential and/or faster progress, but it requires a multidisciplinary team, including, nowadays, molecular biologists. It also needs suitable controlled and field environments and substantial long-term support. All this may no longer be available in the public sector, at least at a single location.