Imagine the larva enclosed in the big ball which the copious quantity of food placed at its disposal might well supply. Once saturated with moisture by a shower, this sphere would retain it stubbornly, for its form is that of least evaporation and of least contact with the sun-warmed soil. In vain, within twenty-four hours, will the surface layer of the ground be restored to its normal coolness: the globular mass will retain its excess of water, for lack of adequate contact with the sun- and air-dried earth. In the too-humid and too-thick recess, the provisions will go musty; the heat from outside will be inopportune, as will the air; and the larva will derive little advantage from this late autumn sun, whose tardy rays ought to ripen it to perfection and give it the necessary vigour to brave the trials of winter. What was a good quality in July, when it was necessary to guard against excessive dryness, becomes a bad one in October, when excessive damp is to be avoided. The cylinder is therefore substituted for the sphere. The new shape, with its exaggerated length, fulfils the converse condition of that beloved by the pill-makers: here, with a similar volume, the surface is developed to its extreme limits. Is there a reason for this complete change? No doubt; and I seem to perceive it. Now that dryness is no longer to be feared, will not this kind of shape, with its large surface, enable the mass of foodstuff to get rid of its superfluous moisture more readily? Should it rain, its wide area certainly will make it liable to more rapid saturation; but also, when the fine weather returns, the surplus water will soon disappear thanks to the extensive contact with a quickly-drained soil. Let us conclude by enquiring how the roly-poly is manufactured. To watch the performance in the fields appears to me a very difficult, not to say impracticable undertaking. With my cages, success is certain, provided that we exercise a little patience and dexterity. I let down the board which keeps the artificial soil in place at the back. This now reveals its vertical surface, which I explore bit by bit with the point of a knife until I strike a burrow. If the operation be cautiously conducted, without the disturbance due to an ill-calculated landslip, the labourers are discovered at their toil, paralysed, it is true, by the sudden flood of light and as it were petrified in the attitude of work. The arrangement of the workshop and the materials, the position and posture of the workers enable us easily to reconstruct the scene, though it be abruptly suspended and not renewed so long as our inspection lasts.