species. On such assumptions it is, of course, impossible to understand how the workers can have come by the obviously adaptive and exquisitely correlated characters, which they are unable to transmit. It will be remembered that neo-Darwinians and neo-Lamarckians, in the persons of Weismann and Herbert Spencer, locked horns over this matter some years ago. Both in this and in many similar discussions, the very premises which both parties accepted are unwarranted. In the first place, it is now known that workers readily become fertile when well fed and that they can and often do produce normal young from unfecundated eggs. Although these young are usually, if not always, males, it is evident that these males, through the eggs which they fertilize, can transmit the characters of their worker mothers to succeeding generations of queens and workers. Thus the congenital, and perhaps even the acquired, characters of the worker are not necessarily lost, but can be gathered up into the germ-plasma of the species. In the second place, most, if not all of the characters of the worker are not qualitatively but only quantitatively different from those of the queen. In other words, the worker does not differ from the queen as a mutant, but as a fluctuating variation, which has been produced by imperfect or irregular feeding during its larval stages. This is true alike of morphological, physiological and psychological characters. Even when the queen fails to manifest the worker instincts, we are not justified in doubting her ability to do so under the proper conditions.
The hitherto unsuspected capacity of the queen ant is beautifully illustrated by another set of facts, which at the same time show the close connection between adaptive behavior and regulation, or regeneration. Under normal conditions the queen, after rearing a brood of workers, no longer takes part in the 'muck and muddle of childraising' but seems to be as indifferent to the young of her species as some women who have brought up large families. If, however, the firstling brood of workers be removed and the queen isolated, she forthwith begins to bring up another brood, precisely as in the first instance, provided her body still contains sufficient food-tissue. She thus regenerates the lost part of her colony, just as a mutilated earthworm regenerates its lost segments. In the ant the absence of workers acts as a stimulus to restore the colony, just as the absence of segments leads the earthworm to complete its body.
The regulatory activities of the queen ant are, of course, highly adaptive and hence evidence of the variability which is so clearly manifested in the physical structure of these insects. There is no contradiction in the coexistence of such variability with the very stable character of certain instincts like those to which I have called attention, for an organism may be extremely plastic in some of its activities and rigidly conservative in others. It is evident that the remarkable vari-
