By the introduction of a bundle of horsehair or catgut, passed through the cannula so as to project into the ventricular cavity, and, after the withdrawal of the cannula, tucked, with respect to the proximal ends, into the subdural, extra-dural, or subaponeurotic spaces, it is obvious that drainage may be established between the ventricles and the other regions. Experience showed, however, that drainage into the subaponeurotic space usually converted the condition of hydrocephalus into one of cephalocele (see Fig. 27), the fluid collecting as a localized fluid tumour over the region of exploration, whilst extra-dural drains did not permit of sufficiently rapid reabsorption of fluid. Subdural drainage gave the best results, the cerebro-spinal fluid being brought into relation with the pia-arachnoid meshwork of vessels. It would, of course, be infinitely preferable if the ventricular fluid could be brought into direct relation with the veins of the _subarachnoid_ space, for the cerebro-spinal tension and venous pressure are equal, and all excess of cerebro-spinal fluid would be absorbed as soon as it is formed. This course is, however, impossible to carry out. We have, therefore, to rest content with less direct contact, drainage into the _subdural_ space. This ventricular-subdural drainage, as obtained by horsehair, catgut, and silk, apparently leads to but temporary benefit, probably owing to falling together of the brain substance and obliteration of the adventitious passage. Silver tubes and bone tubes have been utilized, but the results are sometimes disappointing. In one of my recent cases the two halves of a bone tube were utilized. The tube was cut across in an oblique manner at about its centre, the two parts set at right angles to one another and sewn together with silk. One arm is introduced into the ventricle, the other tucked underneath the dura mater. The child improved considerably, but the method is not altogether satisfactory and by no means easy of application. In another case I utilized strands of silver wire. The depth of brain-tissue necessary to reach the ventricular cavity was measured, and two or three strands of wire introduced so as to project well into that space, then steadied with forceps whilst the proximal ends were bent at right angles to the surface of the brain and tucked underneath the dura mater. The method was unsatisfactory.