Thanks for your question.  It is one I've encountered in the past, and not been able to answer to my own satisfaction!  There is no research on the topic, except for that in hybridization, and the inability of a hybrid to sustain its characteristics through several generations.  As one source reported:

"According to Ellis Fletcher [retired Professor of Horticulture at University of Louisiana], color changes may be produced in some of the offspring and not in others. If you can bear with a complicated explanation, here goes:

Let's assume we have a bed of both solid-red and solid-white amaryllis bulbs. All the red amaryllis — the original bulbs as well as any bulblets developing from the original bulb — would forever remain red. All white amaryllis — the orginal bulbs as well as any bulblets from the original bulb — would forever remain white.

In other words, parents (bulbs or bulblets forming from the parent) would remain the same, but the children (seedlings) would contain the features of both parents and could have features (different colors, sizes, shapes, growth rate, time of blooming, etc.) of either or both parents.

If these flowers were fertilized by the pollen of a different color. The expression of this exchange of gametes would be seen in the resulting seedlings.

Vegetative or asexual propagation (bulbs or any technique not involving seed) would produce offspring identical to the parent. Sexual or seed propagation would produce offspring that could be very similar to significantly different from the plant upon which the seed was produced."

This, of course, assumes two bulbs not (as I think you indicate here) one bulb that was the same one in subsequent years.  It may just be a mystery; it may be the influence of variable amounts of sun from year to year, or ?  I'll keep looking and let you know if I come up with anything more substantive.  Thanks for sending it in!