THE PREDICTION OF NATURAL PHENOMENA. 541 



In many parts of Europe the saying is common, " A western wind 

 brings rain," and it is undeniable that there is a certain connection 

 between rain-storms and a wind from the west. Occasionally, how- 

 ever, rain will put in an appearance from whatever quarter the wind 

 may be blowing. But, as rain is experienced particularly often during 

 westerly winds, the statement above " a western wind brings rain " 

 may, with some slight reserve, be permissible. 



Almost every one owns a barometer. If, in the summer-time, some 

 little excursion is planned, the mercury in the tube is watched with 

 anxious eye to see whether it will rise, for this is generally regarded 

 as predicting clear and dry weather, whereas the sinking of the mer- 

 cury points to rain and storm. Until quite recently the true relation 

 between these phenomena was not known to science, and yet it could 

 be safely assumed that there was, in some way, a connection between 

 them. Thus, Otto von Guericke, the inventor of the air-pump, in the 

 year 1660 correctly predicted a storm from a considerable depression 

 that he observed in the water-column of his immense barometer, which 

 measured nineteen Magdeburg ells in length. 



If we look back for a moment on these predictions in the field of 

 natural science, we must admit that they can not lay claim to any 

 great degree of accuracy. In Europe, as well as elsewhere, a westerly 

 wind is often accompanied by fair weather ; and sometimes it will 

 rain, although the mercury has risen in the barometer. 



Predictions like these may hence be made, but with a certain 

 amount of reserve. Some of them, however, are predictions of a 

 superior order to be discussed hereafter as in many of them the 

 accompanying conditions are studied, and hypotheses as to the phe- 

 nomena observed are formulated and discussed in connection with 

 others. The generalizations considered thus far may only be looked 

 upon as probable ; there are, however, instances in which the prob- 

 ability can be more fully depended on, although, in these too, the sim- 

 ple method of enumeration is employed. 



A case where the probability of the prediction borders on certainty 

 is found in chemistry in enumerating the properties of chemically pure 

 substances. If we say, "All globules of mercury have a grayish-white 

 color, a metallic luster, are opaque," etc., these are assertions true toith- 

 out exception as far as mercury in the perfectly pure state is concerned. 

 If we are dealing with a substance that is chemically pure, we can pre- 

 dict with certainty that it is endowed with certain properties. In fact, 

 we may state it as a natural law that, under all circumstances, sub- 

 stances which have been recognized as identical by the comparison of 

 a series of properties will exhibit other series of like properties. "We 

 will call this the law of coexistence of like properties, or, to be brief, 

 the law of coexistence. 



The counterpart of this is the law of universal causation. This 

 law, so important a one for predictions, may be thus expressed : under 



