The apertures connecting auricles and ventricles are extremely wide, allowing the contents of the former to be emptied into the latter almost instantaneously. If we attempted to make a pump fulfilling this condition, we should find that it failed in several respects. In the first place, the rush of fluid from the one chamber into the other would press the flaps of the valves back against the wall of the second chamber. They would cling to the wall, and would not float up quickly into place when the second chamber was squeezed. Let us call the two chambers A and V for brevity’s sake. When V contracted, some of the fluid would be thrown back into A, because, the resistance in that direction being lower than the resistance offered by the column of fluid above the pump (the resistance in the aorta is very high), the contents of V would rush past the margins of the A-V valve. This would happen even though its flaps were not pressed back against the wall. Further, at the height of contraction the membranous valve would bulge backwards into A, making a cup towards V which V could not empty. In the heart these difficulties have been overcome. The tricuspid valve, which separates the right auricle from the right ventricle, has three flaps. The mitral valve, on the left side of the heart, has but two. The flaps are composed of tough membrane, but are comparatively thin. The following direction for deciding at an autopsy whether or not they were healthy at the time of death was given many years ago by a surgeon of repute: “You ought to be able to see the dirt under your thumbnail when you place it beneath one of the flaps.” Surgery has improved in cleanliness as well as in other ways; indeed, the possibility of advance has been due to the recognition of the need for transcendental cleanliness. But this is a digression. The margins of the flaps are crenulated. Threads—chordæ tendineæ—are attached to them like the stay-ropes of a tent. At their other end these tendons are attached to the musculi papillares already mentioned. The bunch of tendons from each papillary muscle spreads, to be inserted into the contiguous margins of two flaps. We have mentioned some of the difficulties which have been overcome in the construction of the pump. (1) The flaps do not flatten back against the wall of the ventricle during systole of the auricle. It must be remembered that during diastole of both chambers blood is flowing through the auricle into the ventricle. The latter being partly filled before systole of the auricle commences, the flaps are floated up. This is greatly favoured by the form of the inner wall of the ventricle. It is not flat, but raised in pillars—columnæ carneæ. The spaces between these pillars cause backwash currents, which lift the flaps and help to bring them into apposition as soon as systole of the ventricle commences. (2) No blood which has entered the ventricle is thrown back into the auricle.