Another indication of sensory control is found in the repair-work performed by animals endowed with the constructive instinct. C. F. Schroeder, for instance, experimenting on certain caterpillars, found that they repaired their weaving, whenever it was disturbed by the experimenter. Fabre, too, discovered that a Mason-bee would plaster up holes or clefts marring the integrity of its cell, provided that the bee was actually engaged in the process of plastering at the time, and provided that the experimenter inflicted the damage at the level, and within the area, of the construction work on which the bee was then engaged. In a word, if the damage inflicted could be repaired by a simple continuation or extension of its actual work of the moment, the bee was able to cope with the emergency. There are other ways, too, in which the animal adapts its constructive instincts to external circumstance. Fabre tells us that the Bramble-bee _Osmia_, which builds a train of partitioned cells in snail shells or in hollow reeds, will victual first and then plaster in a partition, if the reed be narrow, but will first plaster a partition, and then introduce honey and pollen through a hole left unclosed in the partition, whenever the reed is of greater diameter. This reversal of the procedure according to the exigencies of the external situation does not suggest the chain-reflex of Loeb. (Cf. “The Bramble-Bee,” pp. 214-217.) Another kind of adaptation of instinct to external circumstances consists in the economical omission of the initial step of a serial construction, in cases where the environmental conditions provide a ready-made equivalent. “The silkworm,” says Driesch, “is said not to form its web of silk if it is cultivated in a box containing tulle, and some species of bees which normally construct tunnels do not do so if they find one ready made in the ground, they then only perform their second instinctive act: separating the tunnel into single cells.” (“Science & Phil. of the Organism,” vol. II, p. 47.)