Milk for yogurt manufacture is subjected to high heat treatment to denature whey proteins. Low milk pH values (<= 6.5) at heating result in most denatured whey proteins becoming associated with casein micelles, whereas high milk pH values (>= 7.0) at heating result in the formation of mostly soluble (nonmicellar) denatured whey protein complexes. There are conflicting reports on the relative importance of soluble and casein-bound whey protein aggregates on the properties of acid gels. Prior studies investigating the effect of pH of milk at heating used model gels in which milk was acidified by glucono-delta-lactone; in this study, we prepared yogurt gels using commercial starter cultures. Model acid gels can have very different texture and physical properties from those made by fermentation with starter cultures. In this study, we investigated the effects of different pH values of milk at heating on the rheological, light backscatter, and microstructural properties of yogurt gels. Reconstituted skim milk was adjusted to pH values 6.2, 6.7, and 7.2 and heated at 85 degrees C for 30 min. A portion of the heated milk samples was readjusted back to pH 6.7 after heating. Milks were inoculated with 3% (wt/wt) yogurt starter culture and incubated at 40 degrees C until pH 4.6. Gel formation was monitored using dynamic oscillatory rheology, and parameters measured included the storage modulus (G') and loss tangent (LT) values. Light-backscattering properties, such as the backscatter ratio (R) and the first derivative of light backscatter ratio (R'), were also monitored during fermentation. Fluorescence microscopy was used to observe gel microstructure. The G' values at pH 4.6 were highest in gels made from milk heated at pH 6.7 and lowest in milk heated at pH 6.2, with or without pH adjustment after heating. The G' values at pH 4.6 were lower in samples after adjustment back to pH 6.7 after heating. No maximum in the LT parameter was observed during gelation for yogurts made from milk heated at pH 6.2; a maximum in LT was observed at pH similar to 4.8 for samples heated at pH 6.7 or 7.2, with or without pH adjustment after heating. Higher R-values were observed with an increase in pH of heating, with or without pH adjustment after heating. The sample heated at pH 6.2 had only one major peak in its R' profile during acidification, whereas samples heated at pH 6.7 and 7.2 had 2 large peaks. The lack of a maximum in LT parameter and the presence of a single peak in the R' profile for the samples heated at pH 6.2 were likely due to the partial solubilization of insoluble calcium phosphate when milk was acidified to this lower pH value. No clear differences were observed in the microstructures of gels between the different treatments. This study indicates that heating milk at the natural pH (similar to 6.7) created an optimum balance of casein-bound and soluble denatured whey proteins, which resulted in yogurt with the highest gel stiffness.