As we have said, no human eye has ever seen, or will see, an atom; only the mind's eye, or the imagination, sees atoms and molecules, yet the atomic theory of matter rests upon the sure foundation of experimental science. Both the chemist and the physicist are as convinced of the existence of these atoms as they are of the objects we see and touch. The theory "is a necessity to explain the experimental facts of chemical composition." "Through metaphysics first," says Soddy, "then through alchemy and chemistry, through physical and astronomical spectroscopy, lastly through radio-activity, science has slowly groped its way to the atom." The physicists make definite statements about these hypothetical bodies all based upon definite chemical phenomena. Thus Clerk Maxwell assumes that they are spherical, that the spheres are hard and elastic like billiard-balls, that they collide and glance off from one another in the same way, that is, that they collide at their surfaces and not at their centres. Only two of our senses make us acquainted with matter in a state which may be said to approach the atomic--smell and taste. Odors are material emanations, and represent a division of matter into inconceivably small particles. What are the perfumes we smell but emanations, flying atoms or electrons, radiating in all directions, and continuing for a shorter or longer time without any appreciable diminution in bulk or weight of the substances that give them off? How many millions or trillions of times does the rose divide its heart in the perfume it sheds so freely upon the air? The odor of the musk of certain animals lingers under certain conditions for years. The imagination is baffled in trying to conceive of the number and minuteness of the particles which the fox leaves of itself in the snow where its foot was imprinted--so palpable that the scent of a hound can seize upon them hours after the fox has passed! The all but infinite divisibility of matter is proved by every odor that the breeze brings us from field and wood, and by the delicate flavors that the tongue detects in the food we eat and drink. But these emanations and solutions that affect our senses probably do not represent a chemical division of matter; when we smell an apple or a flower, we probably get a real fragment of the apple, or of the flower, and not one or more of its chemical constituents represented by atoms or electrons. A chemical analysis of odors, if it were possible, would probably show the elements in the same state of combination as the substances from which the odors emanated.