which did not vary favourably in this respect perishing with their hives in consequence of competition and leaving no descendants. Now the variations in the queens which caused them to produce better and better workers cannot have been acquired, for queens do not work, and the evolution of workers can therefore have been due only to the accumulation of inborn variations in the successive queens. It is true that the actual anatomical structure and the instincts of the queens and workers, one or both, are to be regarded in great measure as acquired variations, since they result largely from the kind of food supplied to the larvæ, but these acquired variations in the workers cannot be transmitted to offspring, for, as I say, they have no offspring. The very power of acquiring them must have resulted from the accumulation of inborn variations.
The social system and the subdivision of labour among bees (and ants) are so remarkable, that they present perhaps the most extreme examples of evolution to be found in nature, not even excepting that of man. Yet in the face of this we find many writers insisting, in cases as simple as that of the evolution of horns, that acquired variations are factors of the evolution, and doing so when, in other strictly analogous cases, they admit that acquired variations can have had no part. For example, Mr. Herbert Spencer says (Principles of Biology, vol. i. pp. 295-6)—
