The difficulty, however, comes most clearly to the front when we attempt to define the maximum, and that not for an individual or generation only, but for the race. It is not happiness merely, but greatest happiness, that is the utilitarian end. Is there any way, then, of determining how the maximum of happiness is to be obtained for generations whose characters, though inherited from present individuals, may be modified almost indefinitely? The very existence and numbers of these future generations are problematic; and Mill, as is well known, spent much of his energy in trying to convince the present generation to restrict the numbers of the next. Even on the fundamental question as to whether happiness is to be obtained by the restriction of desires or by the satisfaction which leads to their recurrence and increase, no principle can be extracted from utilitarian ethics. The theory of evolution has shown how desires may be uprooted in the character of the race, though they remain to the end in the present individuals; but in each case utilitarianism would require us to sum up and estimate the relative advantages of renunciation and satisfaction,--a problem which the modifiability of human character seems to make impracticable. Thus, even if certain rules of living may be ascertained, and justified by the utilitarian principle, it would seem that the end of greatest happiness for the race of man, or the sentient creation generally, must remain "abstract." There seems no principle through which it may be applied to conduct--no hope of an accurate estimate of results--when the variability of the individual and of social relations is taken into account. [Sidenote: (b) Objection to utilitarianism as a theory of consequences;]