Recognition by the pigeon of stimuli varying in two dimensions.

Researchers taught pigeons to peck only when two visual cues matched a hidden rule.

One cue was color wavelength. The other was line tilt. Birds had to watch both to win food.

The team then tested new mixes of color and tilt to see how the birds reacted.

Pigeons treated the cues as if they multiply, not just add.

A big color change drowned out small tilt shifts. A big tilt change did the same to color.

The birds’ choices followed smooth wavelength curves, but the height of those curves changed when tilt was also varied.

Snapper et al. (1969) first showed pigeons can learn two-part rules, yet they saw form cues win over color. Striefel (1972) adds the why: once a dimension leaps far enough, it hijacks control.

Sailor (1971) mapped how pigeons generalize compound cues. Striefel (1972) goes further by showing the math inside the gradient—dimensions interact multiplicatively, not just side-by-side.

Wilkie (1973) followed one year later and seemed to clash: birds ignored one dimension unless the payoff forced them to watch both. The gap is method. S used wide value ranges that naturally let one cue dominate. M kept ranges narrow and tied food to both cues, proving control can be shared when the contingency demands it.

When you write conditional-discrimination programs, spacing matters. If color, shape, or size jump too far apart, learners may lock onto that leap and miss the second cue. Start with close values and equal payoff for both, then widen the range only after accuracy holds. Check generalization early; a smooth curve that collapses when you tweak the partner cue signals lopsided stimulus control.