A different covariance matrix is relevant for the B main effect because the B effect averages over levels of A, whereas the A effect averages over levels of B. Table 12.9 presents each subject's mean score for noise absent and noise present, where the mean is the average of the three angle scores at that particular level of noise.
References
Maxwell, S. E., Delaney, H. D., \& Kelley, K. (2018). Designing experiments and analyzing data:
A model comparison perspective (3rd ed.). New York, NY: Routledge.