Under favourable growing conditions, the source-sink relationships in grain yield of sorghum were analysed in terms of capacities for net assimilation, of head storage, and of the transport system to move assimilates between source and sink. In four commercial hybrids (Dekalb E57, Pacific Goldfinger, Texas 610SR, and Texas 626) grown at three population densities (20.2, 40.4 and 80.8 plants m-1), the assimilates supply was varied by increasing or decreasing the radiation available per plant (by thinning or shading), the potential gram storage capacity was decreased by spikelet removal, and the transport system was reduced by incision of the culm, all manipulations being performed at anthesis. Decreasing the number of grains increased the size of those remaining in all cultivars at all population densities The degree of Increase was greatest for T626 and T610 and least for E57 Thinning increased the grain size In all cultivars, but only sufficiently to cause a significant increase In gram yield In T610 and T626 Reduction in the assimilate supply by shading decreased the gram size and yield in all cultivars. Reduction In the transport system had no significant effect on gram yield. These results showed that there was surplus capacity for storage and transport In all cultivars In T626, T6L0, and Goldfinger, all post-anthesis assimilate was stored as gram and grams could grow larger The yield was therefore completely source-limited In E57, however, not all port-anthesis assimilate was stored as gram, and these grams showed little capacity to grow larger, which suggests that the yield was partially limited by both source and sink.

In field and glasshouse experiments with grain sorghum (cv. RS610), the assimilate supply was varied by increasing or decreasing radiation and carbon dioxide supply; the potential grain storage capacity was altered by spikelet removal; and the transport system was reduced by incision of the culm. Plants grown at four population densities in the field were manipulated to increase (by removing neighbouring plants) or decrease (by shading) the supply of photosynthates during grain filling. These treatments affected grain size and thus yield. Removal of some of the spikelets at three-quarter anthesis resulted in a significant increase in the size of those grains remaining at maturity. From anthesis onward, a reduction in the capacity of the transport system in the culm had no significant effect on grain yield. These results are interpreted as evidence that grain yield is not limited by the storage capacity of the grain, or by the transport system involved in moving material from the stem to the grain. Treatments which altered the demand for assimilates by the grain, relative to the supply, did not affect net photosynthesis. Dry matter produced in excess of grain requirements accumulated in other plant parts, including the root. Potential grain size was influenced by interspikelet competition operating within 1 week after three-quarter anthesis. *Part II, Aust. J. Agric. Res., 22: 39-47 (1971).

Eleven grain sorghum (Sorghum bicolor L. Moench) hybrids were grown in the field and growth chamber. Leaf blade samples were assayed by in vitro and in vivo procedures for nitrate reductase activity (NRA) to i)termine NRA of the hybrids, ii) examine relationships between NRA and grain and N yields, and iii) determine whether NRA of growth chamber‐grown plants are indicative of those of field‐grown plants. Among hybrids, differences in yields of grain, dry matter, and grain protein and in in vitro NRA were significant though comparatively small. Seven hybrids had similar N input: accumulation ratios while four had lower ratios. This variation in N input:accumulation ratios, precludes the use of NRA, as assayed here, as a criterion for screening grain sorghum cultivars for N accumulation ability. Also, NRA values of young growth chamber‐grown plants were not indicative of those of field‐grown plants.

The relationships between illuminance, leaf resistance, and net photosynthesis were studied in sorghum. Some comparisons were made between varieties chosen on the basis of possible differing environmental adaptation, and with cotton whose photosynthetic response is known to differ from that of sorghum. The after-effects of varying periods and degrees of shading were examined for sorghum. Net photosynthesis of sorghum was not light saturated at 11,000f.c. and increases above intermediate illuminance arises almost entirely from increasing stomatal opening. At lower intensities other factors become important, but leaf resistance still exerts the major control. In contrast, leaf resistance in cotton was much lower at all illuminances, and was relatively little affected by changing illuminance. Light saturation of cotton at 5000 f.c. was associated with constant leaf resistance, while at lower intensities, declining net photosynthesis must depend mainly on factors other than stomatal closure. Variation between four sorghum varieties in response to illumination was attributed solely to differences in leaf resistance. Stomata opened slowly after shading at below 3000 f.c., and at a rate independent of the duration of shading. Shading in the afternoon had greater and more persistent effects. Net photosynthesis was affected according to stomatal aperture.

Seedlings of 41 cultivars of grain sorghum [Sorghum bicolor (L.) Moench] including 30 hybrids and all but two of their parents were grown in a growth chamber; harvested 10, 14, and 18 days after planting; and assayed for hydrolizable cyanide and in vitro and in vivo nitrate reductase activity (NRA). The genotype X age variance component was essentially equal to the genetic variance for each character, which suggest that genotype X environment interactions are major components of the phenotypic variance. Cyanide level was highest in 10‐day‐old plants and dropped sharply in 14‐day‐old plants. Broadsense heritability of hydrolizable cyanide was high (0.89 to 0.95) for all three ages of plants. Narrow‐sense heritability indicated the cyanide content of 10‐day‐old plants could be changed 0.76 units for each unit change in the midparental value, but expected progress decreased markedly for 18‐day‐old plants. Broad‐sense heritability of in vitro NRA was high at all ages, and narrow‐sense heritability was moderate to low. Thus, significant genetic advance could be expected from selection among pure lines and hybrids, but nonadditive genetic variance would reduce progress in hybrids based on selection of their parents to a moderate level. In vivo NRA was only 20% of in vitro activity. Heritabilities, in general, also were lower for in vivo than for in vitro NRA.