When July arrives, this doubly-tenanted house becomes the scene of a tragic conflict. Those below, on attaining the adult state, burst their swaddling-bands, demolish their resin partitions, pass through the gravel barricade and try to release themselves; those above, larvae still or budding pupae, prisoners in their shells until the following spring, completely block the way. To force a passage from the far-end of those catacombs is beyond the strength of the Resin-bee, already weakened by the effort of breaking out of her own nest. A few of the Osmia's partitions are damaged, a few cocoons receive slight injuries; and then, worn out with vain struggles, the captives abandon hope and perish behind the impregnable wall of earth. And with them perish also certain parasites, even less fit for the prodigious work of clearance: Zonites and Chryses (Chrysis flammea), of whom the first are consumers of provisions and the second of grubs. This lamentable ending of the Resin-bee, buried alive under the Osmia's walls, is not a rare accident to be passed over in silence or mentioned in a few words; on the contrary, it happens very often; and its frequency suggests this thought: the school which sees in instinct an acquired habit treats the slightest favourable occurrence in the course of animal industry as the starting-point of an improvement which, transmitted by heredity and becoming in time more and more accentuated, at last grows into a settled characteristic common to the whole race. There is, it is true, a total absence of positive proofs in support of this theory; but it is stated with a wealth of hypothesis that leaves a thousand loopholes: 'Granting that...Supposing that...It may be...nothing need prevent us from believing... It is quite possible...' Thus argued the master; and the disciples have not yet hit upon anything better. 'If the sky were to fall,' said Rabelais, 'the larks would all be caught.' Yes, but the sky stays up; and the larks go on flying. 'If things happened in such and such a way,' says our friend, 'instinct may have undergone variations and modifications.' Yes, but are you quite sure that things happened as you say? I banish the word 'if' from my vocabulary. I suppose nothing, I take nothing for granted; I pluck the brutal fact, the only thing that can be trusted; I record it and then ask myself what conclusion rests upon its solid framework. From the fact which I have related we may draw the following inference: