My goal was to determine how the pine larvae are able to maintain their orderly processions. First off, I found that the caterpillars left behind a persistent trail. I did this by marking the precise course of a procession with pins and then releasing single, uninitiated caterpillars along the pathway. That simple test revealed that individual caterpillars followed the procession when released as long as a full day after the procession had passed.
Next I tested Fabre’s hypothesis that the strands of silk the caterpillars lay down as they march are the basis for the trail: I temporarily plugged the tips of the silk-spinning apparatus, the spinneret, of each member of a group of caterpillars. That had no effect on their ability to line up and follow one another, or on the ability of a caterpillar introduced to the trail hours later to follow its path. The trail was an invisible one.
Further investigation revealed, as I had suspected, that processionaries mark their pathway with a pheromone secreted from the underside of their abdomens. When I sealed the suspected site of pheromone secretion, as well as the spinnerets, the caterpillars did not leave a persistent trail that a lone latecomer could follow. Nevertheless, they still formed and maintained processions. That suggested tactile stimulation—caterpillars brushing each other from behind—was the ultimate basis for the formation and maintenance of processions. To test that, I had to sacrifice a caterpillar and pull its eviscerated skin over the end of small stick, making a model of a caterpillar. With the model, I found I could lead a procession in any direction by moving it slowly in front of the leader. Moreover, I was able to do the same with a model made from a completely unrelated species of caterpillar.
If tactile stimuli alone are adequate to allow the caterpillars to march in line, what then are the functions of the silk pathway and the trail pheromone? I concluded that the sticky silk, bonding to the smooth branches of the host tree, gives the caterpillars’ feet extra traction. And the persistent pheromone serves to guide the caterpillars back to the nest after their nighttime feeding foray (and also enables laggards to catch up). The caterpillars discriminate between old and new pheromone trails and between trails made by different numbers of individuals. That facilitates efficient and coordinated movement between the nest and distant feeding sites.
During their final over-the-ground procession, however, the caterpillars rely solely on tactile stimuli to stay together. (The leader has no foreknowledge of where to go, but may use the Sun for direction.) I learned this by rapidly removing single, centrally located caterpillars from their places in line, dividing processions in two. The new leader of a rear contingent would swing its body in a frenetic search for the caterpillar that had been snatched away, then end up striking off independently in a new direction. None succeeded in bridging the gap I had created. Either the caterpillars do not lay down a pheromone trail during these terminal marches or the rough terrain renders it too spotty to be a useful guide.
And what of Fabre’s caterpillars that circled endlessly around the rim of a pot? My suspicion was that the caterpillars were not adhering blindly to their instinct to follow each other, but were in effect physically trapped on the rim. I tested this by arranging for caterpillars to make a circular trail on a flat surface rather than on the rim of a pot. I placed seventeen inside a circular arena on a tabletop. The caterpillars soon arranged themselves against the wall of the arena and commenced marching in an unbroken, head-to-tail circular procession. After half an hour I removed the arena, freeing the caterpillars from any physical constraint. If Fabre’s interpretation had been correct, the caterpillars should have continued to circle for days, or until they grew exhausted. But in fact, they continued in a circle for a short time (on average in my experiments, only two minutes), then marched off in a straight line. When I tried the same experiment with caterpillars younger than those used by Fabre, I learned that the young ones are more dependent on the trails. Their circular processions lasted an average of two and a half hours. One group circled continuously for twelve hours—a remarkable performance, but far short of the seven days.