We see men using as clubs only sticks of a certain length or weight

Donald Redfield Griffin, Animal thinking, Harvard University Press, 1984, p.98-101.

Artefacts and Templates

Caddis Fly Cases

The behavior of another group of aquatic animals that construct protective coverings is more suggestive of intentional action. These are the larvae of caddis flies, which are quite abundant in fresh water streams and ponds. Superficially the larvae are rather like the familiar caterpillars that metamorphose into moths and butterflies. Like caterpillars, they have versatile mouthparts for cutting up particles of vegetation or capturing smaller aquatic animals (Wiggins, 1977). But they develop from eggs laid in the water by the winged female caddis flies, which emerge and mate after a long aquatic larval period.

The larvae of almost all species of caddis flies cover their bodies with particles of sand, bits of leaves, or other available particles, cemented together by silk secreted from glands on the head. In the early larval stages, the animal uses small, homogeneous particles to form a roughly cylindrical case, often with a single particle closing the front of the case. These cases almost certainly protect the otherwise vulnerable larvae from predation by small fishes or predatory larvae of insects such as dragonflies.

The caddis fly case is not totally impervious : a hole at the posterior end allows feces to pass out, and water circulates freely through the case, so the larva’s gills can extract oxygen. When the larva moves about, it pushes its head and the thoracic segments bearing the six legs out through the front opening; small hooklike projections on the abdominal segments hold the case close to the body.

The are several dozen genera of caddis flies throughout the world, and their larval cases are enormously varied and sufficiently characteristic that they are often identified more easily by their cases than by the structure of the larva itself.

The larvae are somewhat selective about the materials used in case construction, within the limits of what is available. Many cases consist of grains of sand or mud or other inert material, but some species cut pieces from the leaves of aquatic plants and a few use the discarded shells of tiny aquatic snails. Some species build structures that serve to capture prey from the flowing water; the case may simply be enlarged at the upstream end, or the larvae may build nets of parallel strands of silk that strain out minute aquatic plants and animals. In the few species that have been studied carefully in the laboratory, the larva can be seen to pick up a variety of objects with its legs, feel them with its mouthparts, and retain only those that are suitable. Those that use leaves cut them into pieces of the right size without making wasteful trials of inappropriate objects. As the larvae grow, their cases must be enlarged.

Sometimes they take over empty cases abandoned by the original builders.

The animal will maintain a particular pattern of case structure, and if an experimenter removes portions of the case, the larva will replace them with pieces of similar size and shape.

Is it possible that even a caddis fly larva has some faint inkling of what it is doing ? Most biologists firmly believe that creatures as simple as insect larvae operate entirely by genetically programmed reflexes and cannot possibly think about anything. But if one spells out the genetic programming that would account for all features of caddis fly case construction, it soon becomes a rather imposing list.

Cecropia caterpillars have been studied thoroughly enough to show that the silk cocoons they construct are formed by means of a few relatively simple behavior patterns involving flexion of the body and emission of silk from glands in a certain phase of the movement pattern (Van der Kloot and Williams, 1953).

Behavior as complex as the caddis fly larvae’s construction of cases and food- catching nets has not been studied in equal detail, but most scientific students of insects argue by analogy that caddis fly cases must result from similar, though necessarily somewhat more involved, reflexes. When a larva cuts a piece of leaf, the dimensions of the piece seem to be related to the size of the animal’s head and anterior appendages. This tempts the reduction to infer that the larva has not made a conscious selection, but that the size of its mouthparts determines the size of the leaf fragment or of the pebbles that is picked up. But if we see men using as clubs only sticks of a certain length or weight, we do not deny that they have thought about their selections.

Detailed studies by Hansell (1968) illustrate the degree of selectivity in the construction of caddis fly cases. One species, Silo pallipes, begins larval life by constructing a simple tube of sand grains cemented together. After each of its five instars the animal sheds its external chitinous skeleton and grows a larger one. Toward the end of the first instar the Silo adds to its cylindrical case, composed of half-millimeter particles, two larger sand grains at the sides of the frontr end. During the second instar it adds two more, larger particles, and through the third, fourth, and fifth instars it follows the same pattern, each time selecting larger grains of sand. By the end of the fifth instar the animal and its case are about ten milimeters long, and the larger, anterior grains of sand are two to five millimeters in size. During all this growth the larva emarges its cylindrical case by adding smaller particles.

Even this relatively simple creature conforms to definite structural patterns when selecting particles for its case. Hansell (1972) studied another species, Lepidostoma hirtum, that cuts panels from leaves to construct a house with a floor , roof, and two sides, all formed from approximately rectangular pieces of leaf one to two millimeters in size held together by secreted silk. The structure is strengthened by the staggered arrangement of the pieces; each joint between two side plates intersects with the middle of a roof plate, and vice versa. That this staggered arrangement is not wholly accidental was shown when Hansell modified the houses. If he cut away the front end to give the structure a continuous smooth front edge, the larva would cut leaves into different shapes than normal and glue them into place, restoring the staggered arrangement.

According to most biologists, this compensatory hehavior can be explained by postulating yet another feature of the genetically programmed behavior pattern. Presumably under natural conditions a house may be damaged, perhaps nibbled by a fish but not totally destroyed. Supposedly, then, for each type of potential damage, the larva’s central nervous system is prepared genetically with instructions for repair. I have discussed the shelter-building behavior of caddis flies in some detail because it poses a very general problem. If we find that an animal constructs a useful and effective structure , how far can we reasonably go in inferring that the animal thinks about what it is doing? A student who suggests such a notion will be firmly corrected by his science teacher and admonished never to mention such an idea again, lest he be judged unscientific. The student is assured that insect larvae operate solely by means of a few fixed reflexes, which, activated sequentially, produce the finished structure. One must postulate that various stimuli activate the reflexes in an appropriate fashion. The larva from whose case a section has been removed does not pick up any old sand grain but feels about until it locates a piece of the right size and shape, or cuts a suitable piece from a bit of leaf. In a very general sense scientists feel that construction of shelters by animals as simple as caddis fly larvae is similar to the morphological development of complex body parts — except that the actions of appendages or jaws in shelter building are visible, whereas the poorly understood mechanisms that guide anatomical development and cause cells to differentiate into kidneys, legs, gills, and brain are not accessible to our scrutiny.


Donald Redfield Griffin, Animal thinking, Harvard University Press, 1984, p.98-101.