466 THE AMERICAN JOURNAL OF SOCIOLOGY

Staff. Both the selection and training of assistants is best when controlled by the governing head of each reformatory. The civil-service system wards off some improper demands for appointments but at the same time restricts the range of selection and hinders prompt si f ting-out of the unfit. Its serviceableness, however, preponderates.

So delicate and easily disturbed is the generative reformative process that outsiders — ^the would-be special philanthropists, pro- fessional religious revivalists, advertising Salvationists — should generally be excluded; or if at all admitted to any participation, their ministrations should, under the direction of the governor, be made to fit into the established culture course. Even a resident official chaplain may inadvertently interfere with the germination of reformations. I have found the resident chaplain to be less desirable for religious ministrations than an itinerant service. One mind, and that the mind of the resident reformatory gov- ernor, must have and hold and yield every operating agency — impel, steady, and direct the whole and every item of the pro- cedure. Such completeness of control requires an exacting and strenuous disciplinary regime which for effectiveness must in- clude the principle and exercise of coercion.

A majority of prisoners instinctively respond to the inherent persuasion of the combined agencies; and of those who do not a majority readily respond to the moral coerciveness of the agencies. Some, only a small ratio, do not respond at first, except to some form of corporal coercion — some bodily inconvenience and discomfort. These, the irresponsive, who, for the good of the prison community and for the public safety most need refor- mation, should not be neglected nor relegated to incorrigibility until every possible effort has unavailingly been made for their recovery. The advantages proffered are, naturally, not appre- ciated until availed of and enjoyed. Some cannot adopt and carry into execution measures calculated for their own good without the intervention of coercion. Adjustment to environ- ment, even if it is compulsory, leads from the avoidance of bodily risks to the avoidance of social risks and thus to non-criminal

habits, which, when duly formed, no longer need the prop of