According to the mood-congruency hypothesis, happy material is learned better in happy moods, and sad material is learned better in sad moods. One explanation of the effect is provided through the associative network model of mood; moods activate concepts with similar valence, which in turn activate still other, similarly valenced material. The greater density of associations that are available when a mood is activated renders mood-congruent material easier to learn and remember. Certain controversies exist concerning the effect; for example (a) it may be easily alterable-or even caused-by instructions, (b) it may be related to mood strength, and (c) it may be moderated by personality variables. One hundred and ninety-six participants went through mood-congruency procedures. Mood congruency appeared, was unrelated to various instructional manipulations, unrelated to reported mood strength, and mostly unrelated to personality variables. Such a pattern suggested a spreading-activation explanation, perhaps due more to cognitive priming than mood per se. © 1990.