While Borelli and his successors are correct as to the wedge-action of the wing, they have given an erroneous interpretation of the manner in which the wedge is produced. Thus Borelli states that when the wings descend their posterior margins ascend, the two wings forming a cone whose base is represented by _c b e_ of fig. 113, p. 220; its apex being represented by _a f_ of the same figure. The base of Borelli’s cone, it will be observed, is inclined forwards in the direction of the head of the bird. Now this is just the opposite of what ought to be. Instead of the two wings forming one cone, the base of which is directed _forwards_, each wing of itself forms two cones, the bases of which are directed _backwards_ and outwards, as shown at fig. 116. [Illustration: FIG. 116.] In this figure the action of the wing is compared to the sculling of an oar, to which it bears a considerable resemblance.[116] The one cone, viz., that with its base directed outwards, is represented at _x b d_. This cone corresponds to the area mapped out by the tip of the wing in the process of _elevating_. The second cone, viz., that with its base directed backwards, is represented at _q p n_. This cone corresponds to the area mapped out by the posterior margin of the wing in the process of _propelling_. The two cones are produced in virtue of the wing rotating on its root and along its anterior margin as it ascends and descends (fig. 80, p. 149; fig. 83, p. 158). The present figure (116) shows the double twisting action of the wing, the tip describing the figure-of-8 indicated at _b e f g h d i j k l_; the posterior margin describing the figure-of-8 indicated at _p r n_. It is in this manner the cross pulsation or wave referred to at p. 148 is produced. To represent the action of the wing the sculling oar (_a b_, _x s_, _c d_) must have a small scull (_m n_, _q r_, _o p_) working at right angles to it. This follows because the wing has to elevate as well as propel; the oar of a boat when employed as a scull only propelling. In order to elevate more effectually, the oars formed by the wings are made to oscillate on a level with and under the volant animal rather than above it; the posterior margins of the wings being made to oscillate on a level with and below the anterior margins (pp. 150, 151). [116] In sculling strictly speaking, it is the upper surface of the oar which is most effective; whereas in flying it is the under.