Under these conditions not a male arrived; not one, though the warmth and quiet of the evening were propitious. Whatever its nature, whether of glass, metal, card, or wood, the closed receptacle was evidently an insuperable obstacle to the warning effluvia. A layer of cotton-wool two fingers in thickness had the same result. I placed the female in a large glass jar, and laced a piece of thin cotton batting over the mouth for a cover; this again guarded the secret of my laboratory. Not a male appeared. But when I placed the females in boxes which were imperfectly closed, or which had chinks in their sides, or even hid them in a drawer or a cupboard, I found the males arrived in numbers as great as when the object of their search lay in the cage of open wire-work freely exposed on a table. I have a vivid memory of one evening when the recluse was hidden in a hat-box at the bottom of a wall-cupboard. The arrivals went straight to the closed doors, and beat them with their wings, _toc-toc_, trying to enter. Wandering pilgrims, come from I know not where, across fields and meadows, they knew perfectly what was behind the doors of the cupboard. So we must abandon the idea that the butterfly has any means of communication comparable to our wireless telegraphy, as any kind of screen, whether a good or a bad conductor, completely stops the signals of the female. To give them free passage and allow them to penetrate to a distance one condition is indispensable: the enclosure in which the captive is confined must not be hermetically sealed; there must be a communication between it and the outer air. This again points to the probability of an odour, although this is contradicted by my experiment with the naphthaline. My cocoons were all hatched, and the problem was still obscure. Should I begin all over again in the fourth year? I did not do so, for the reason that it is difficult to observe a nocturnal butterfly if one wishes to follow it in all its intimate actions. The lover needs no light to attain his ends; but my imperfect human vision cannot penetrate the darkness. I should require a candle at least, and a candle would be constantly extinguished by the revolving swarm. A lantern would obviate these eclipses, but its doubtful light, interspersed with heavy shadows, by no means commends it to the scruples of an observer, who must see, and see well. Moreover, the light of a lamp diverts the butterflies from their object, distracts them from their affairs, and seriously compromises the success of the observer. The moment they enter, they rush frantically at the flame, singe their down, and thereupon, terrified by the heat, are of no profit to the observer. If, instead of being roasted, they are held at a distance by an envelope of glass, they press as closely as they can to the flame, and remain motionless, hypnotised. [Illustration: THE GREAT PEACOCK MOTH. THE PILGRIMS DIVERTED BY THE LIGHT OF A LAMP.]