hour to 11 p. m., averaging 6½ per cent. below mean. It is probable that both these observations are worthless in view of the small number of instances covered. It is clear that they do not concur; yet, taken separately, each would seem conclusive.
Second. Astronomical predictions are based upon a series of unvarying coincidences, in most cases in harmony with laws whose operations can be tested at any time. If these phenomena were irregular and unclassifiable such predictions would be wholly uncertain; but because they usually coincide,—and when they do not, interfering causes can be traced,—eclipses can be foretold for thousands of years in advance, and discoveries such as those of Uranus and Neptune be made.
Third. Chemistry and cognate sciences also work with fixed phenomena, so that when the most diverse elements are combined and effects observed, formulæ can be deduced by which at all times the same effects can be produced.
Fourth. Many of the most wonderful inventions have been made by seeming accident; for example, photography. But reflection upon the accident reveals the cause; the cause and the effect are seen to be scientific coincidences, and the art with its principles and practice is the result.
Fifth. The performances of jugglers are in harmony with the established methods of nature. The charm of their exploits is in successful concealment of the causes, rapidity of motion, distraction of attention, and shrewdly contrived illusions of the senses.
Sixth. It is essential to remember that so-called "laws of chance" reflect no light on the order of sequence. It may be rendered probable by those laws that a certain event will not occur on the average more than once in a million of times; but this
