Osmia tricornis will lend itself to my experiment better than any other insect on account of its dry honey, or bee-bread, which is largely formed of flowery pollen.  I knead it with the albumen, graduating the dose of the latter so that its weight largely exceeds that of the bee-bread.  Thus I obtain pastes of various degrees of consistency, but all firm enough to support the larva without danger of immersion.  With too fluid a mixture there would be a danger of death by drowning.  Finally, on each cake of albuminous paste I install a larva of medium growth.

This diet is not distasteful; far from it.  The grubs attack it without hesitation and devour it with every appearance of a normal appetite.  Matters could not go better if the food had not been modified according to my recipes.  All is eaten; even the portions which I feared contained an excessive proportion of albumen.  Moreover—­a matter of still greater importance—­the larvae of the Osmia fed in this manner attain their normal growth and spin their cocoons, from which adults issue in the following year.  Despite the albuminous diet the cycle of evolution completes itself without mishap.

What are we to conclude from all this?  I confess I am embarrassed. Omne vivum ex ovo, says the physiologist.  All animals are carnivorous in their first beginnings; they are formed and nourished at the expense of the egg, in which albumen predominates.  The highest, the mammals, adhere to this diet for a considerable time; they live by the maternal milk, rich in casein, another isomer of albumen.  The gramnivorous nestling is fed first upon worms and grubs, which are best adapted to the delicacy of its stomach; many newly born creatures among the lower orders, being immediately left to their own devices, live on animal diet.  In this way the original method of alimentation is continued—­the method which builds flesh out of flesh and makes blood out of blood with no chemical processes but those of simple reconstruction.  In maturity, when the stomach is more robust, a vegetable diet may be adopted, involving a more complex chemistry, although the food itself is more easily obtained.  To milk succeeds fodder; to the worm, seeds and grain; to the dead or paralysed insects of the natal burrow, the nectar of flowers.

Here is a partial explanation of the double system of the Hymenoptera with their carnivorous larvae—­the system of dead or paralysed insects followed by honey.  But here the point of interrogation, already encountered elsewhere, erects itself once again.  Why is the larva of the Osmia, which thrives upon albumen, actually fed upon honey during its early life?  Why is a vegetable diet the rule in the hives of bees from the very commencement, when the other members of the same series live upon animal food?