One session he began the first lecture of the higher class by writing dx⧸dt in the middle of the blackboard, and demanding of each of the ten or a dozen students present, some of them distinguished graduates, what it meant! One student described it as the limiting value of the ratio of the increment of the dependent variable x to the increment of the independent variable t, when the latter increment is made indefinitely small. He retorted, "That's what Todhunter would say!" The others gave various slightly different versions of the same definition. At last he impatiently remarked, "Does nobody know that dx⧸dt means velocity?" Here the physical idea as a whole was before his mind; and he did not reflect that if t denoted time and x distance in any direction, the explanation given by the student did describe velocity with fair accuracy. An embarrassing peculiarity of his mathematical discussions was his tendency, when a difficulty of symbolisation occurred, to completely change the notation. Also he was not uniformly accurate in analytical work; but he more than made up for this by the faculty he had of devising a test of the accuracy of the result and of divining the error which had crept in, if the test was not satisfied. The subjects he treated were always such great branches of mathematics as the theory of the tides--he discussed the tidal phenomena of the English Channel in one course--the general theory of vibrations, Fourier analysis, the theory of waves in water, etc., etc. A very good idea of the manner and matter of his mathematical prelections can be obtained from a perusal of the _Baltimore Lectures_. In the physical laboratory he was both inspiring and distracting. He continually thought of new things to be tried, and interrupted the course of the work with interpolated experiments which often robbed the preceding sequence of operations of their final result. His ideas were on the whole better worked out by a really good corps of students when he was from home, and could only communicate by letter his views on the work set forth in the daily reports which were forwarded to him. He insisted with emphasis that a student who found that a quadrant electrometer would not work well should take it to pieces to ascertain what was the matter. This of course generally resulted in the return of the instrument to White's shop to be put together again and adjusted. But, as he said, there was a cause for every trouble of that kind, and the great thing was to find out at once what it was.