5 per cent. of the speed of the screw.  For the ship F, O is the point of no apparent slip, and the real slip is (5 + 11.4) or 16.4 per cent.  For E, the point of real slip is approximately the same as for F. For B and D, the positions on the curve would be about the same.  The ship B has a higher speed of wake than D, but the screw D has the greater apparent slip.  The influence of the number of blades on the scale for the slip has been neglected.  If this efficiency curve were applicable to full sized screws propelling actual ships, and if the determination of the wakes were beyond question, then we should have a proof that our screws were at or near the maximum efficiency.  But, as we know, from the total propulsive efficiencies, that the screws have high and not widely different efficiencies on these ships, we may argue the other way, and say that there is good reason to consider that at least the upper part of the curve agrees with experience obtained from actual ships.  Now take Fig. 6 and consider the general laws there represented.  Take the speed of the wake as 10 per cent. of the speed of the screw, which is probably an average of widely different conditions, including many single as well as twin screw ships.  Then this curve shows that considerable negative slips mean inefficient screws; that screws may have very different positive slips without any appreciable difference in their efficiencies; and that very large positive slips and inefficient screws may be companions.  For instance, a screw with a large positive slip in smooth water is frequently inefficient at sea against a head wind, which increases the resistance, and necessitates an increase of slip.  I venture to say that these statements, taken in a general manner, are not at variance with experience obtained from the performances of screw ships.  Before it is possible to satisfactorily decide if this curve applies in a general manner to full sized screws propelling ships, we require the results of trials of various ships where the screws are working about the region of no slip.  Model experiments teach that the scale for the slip varies with the design of the screw, and that with a given screw the speed of the wake (which decides the point of no apparent slip) varies with the type of ship and with the position of the screw with respect to the hull.  Remembering these disturbances, it is not improbable that it may be possible to account for or explain what at first sight may appear departures from the curve.  The diameters of the screws in the table are not compared with the diameters given by the method explained by Mr. Froude in his paper last year, for there are differences in the slips, the proportions of blade area to disk, and, to some extent, in the shapes of the blades, which are not taken into account in that method.  Assuming, however, as Mr. Froude does, a constant proportion of blade area to disk, and a uniform pattern of blade, the determination of the diameter for a given set of conditions may, as a rule, be a complete solution of the problem of the design of a screw, but these assumptions do not cover all the necessities of actual practice, which make it extremely desirable to know something about the influence or efficiency of various proportions of blade area to disk, and of the form or distribution of a given area.