The silken filament, which when drawn out appears to be one thread, is composed of two fibres, unwound through the two orifices before described; and these fibres are brought together by means of two hooks, placed within the silk-worm’s mouth for the purpose. The worm rests on its lower extremity throughout the unwinding operation, and employs its mouth and front legs in the task of directing and uniting the two filaments. The filament is not wound in regular concentric circles round the interior surface of the ball, but in spots, going backwards and forwards with a sort of wavy motion. This apparently irregular manner of proceeding is plainly perceptible when the silk is being reeled off the ball; which does not make more than one or two entire revolutions while ten or twelve yards of silk are being transferred to the reel[133]. [133] Mr. Robinet, of Paris, made the following curious calculation on the movements a silk-worm must make in forming a cocoon supposed to contain a thread of 1500 metres. It is known, says Mr. Robinet, that the silk-worm, in forming his cocoon, does not _spin_ the silken filament in concentric circles round the interior surface of the ball, but in a zigzag manner. This it effects by the motions of its head. Now if each one of these motions gives half a centimetre of the silken filament; it follows that the worm must make 300,000 motions of its head to form it; and if the labor requires 72 hours in the performance, the creature makes 100,000 motions every 24 hours, 4,166 per hour, 69 per minute, and a little more than one in a second! At the end of the third or fourth day, the worm will have completed its task; and we have then a _silk cocoon_ (See Figure 3. Plate III.), with the worm imprisoned in its centre; the cocoon being from an inch to an inch and a half long, and of a yellow or orange color.