with the impression that, though they might be very effective watermen, they had but a poor style. Soon after, however, I had occasion to watch Oxford oarsmen at their work, and I found that they row in a style which, without being actually identical with that of the London waterman, resembles it in all essential respects. The moment the oar catches the water, the body is thrown back as in the Cambridge style, but the arms, instead of being kept straight, immediately begin to do their share of the work. The result is that when the body is upright the arms are already bent, and the stroke is finished when the body is very little beyond the perpendicular position. Now let us compare the two strokes theoretically. In each stroke the body does a share of the work, and in the Cambridge stroke the body even seems to do more work than in the Oxford stroke, since it is swayed farther back. In each stroke, again, the arms do a share of the work, but in the Oxford stroke the work of the arms is distributed equally as a help to that of the body, whereas in the Cambridge stroke the work of the arms is all thrown upon the finish of the stroke. At first sight it seems to matter very little in what order the work is done, so long as the same amount of work is done in the same space of time. But here an important consideration has to be attended to. There are two things which the oarsman does in whatever style he rows. He propels the boat along, by pressing the blade of his oar against the water as a fulcrum; but he also propels his oar more or less through the water. If instead of the actual state of things, the boat were to slide along an oiled groove in some solid substance, whose surface was so ridged that the oar could bear upon the ridges without any flexure, then indeed it would matter very little in what way the oar was pulled, so long as it was pulled through a good range in a short space of time. But the actual state of things being different, we have to inquire whether it is not possible that one style of rowing may serve more than another to make the slip of the oar through the water (a dead loss, be it remarked, so far as the propulsion of the boat is concerned) bear too large a proportion to the actual work done by the rower.