Extent to which relative differences affect main versus interaction effects

I am wondering about the following thing. Two (brain area 1 and brain area 2) 2-way (Treatment x Valence) ANOVA were used to examine both interaction and main effects.
In the first analysis (of brain area 1), the interaction effect was significant: F(2, 38) = 4.2, p < .05. The interaction effect indicates that in one of the treatment conditions, valence scores did differ, and in the other treatment conditions, they did not differ significantly. No main effects were found. Means were not provided, but the figure indicates that the maximum difference between these six means was about .2 .
In the second analysis (of brain area 2), the interaction effect was not significant: F (1,19) = 2.1, n.s. A main effect of Valence was significant, F (2,38) = 29.9, p < .001. Means again were not provided, but the figure indicates that the difference between the highest and lowest mean was about .8 .

I wonder if it is possible that the difference in means between treatment conditions were just as big in the second analysis as in the first analysis, but that does not lead to a significant interaction in the second one, because the means have such bigger variance because of the big Valance effect.

I need to review the article, and I sense something is seriously wrong, but I find it very hard to exactly find what it is. Thanks so much in advance!