A logical division of the subject matter of physical geography is into three studies: that of the atmosphere, that of the waters, and that of the earth. In each of these there is room for a great deal of hairsplitting about what belongs to geography and what to geology, geophysics, or meteorology. Physical geography merges into the other natural sciences as human geography merges into history, politics, economics, or ethnology. Even at the present day, when the often futile attempt is being made to delimit the domains of the various sciences ever more definitely, it is impossible to distinguish where one begins and another ends, and it would be foolish to set up hard and fast definitions in dealing with the lore of the ancient and medieval worlds, when natural science was as yet inchoate. METEOROLOGY The ancients were more interested in meteorology[74] than they were in oceanography and physiography (if such terms can be used for their naïve attempts at explaining the features of ocean and land), perhaps because the phenomena of the air make a deeper impression on men than the phenomena of the sea and earth—tides, earthquakes, and volcanoes excepted. Thunder and lightning, comets, rainbows, balls of fire were looked upon as portents, and complex theories were created to explain them and what they were supposed to foretell. But all this type of meteorological lore, however interesting in itself, is, strictly speaking, not geography. On the other hand, there are certain distinctly geographical aspects of the study of the atmosphere as pursued by the Greeks and Romans that deserve our attention.