Let us take the case of odours. One excludes another; they are not associated in the imagination as visual images are in the recollection of a beautiful landscape. One of my correspondents is able to recall at will the scent of pinks; she tried to do so while walking in a wood full of decaying leaves and their smell, but without success; one odour excluded the other. Neither can they be arranged in sequences. I am aware that an English chemist, Piesse, has claimed the ability to class scents in a continuous series, like notes, patchouli corresponding to lower _c_ in the key of F, and civet to the upper _f_ in the key of G—the whole in tones and semitones; but no one, so far as I know, has taken this fancy seriously. In the same way, tastes may be associated with other images, as of hunger (I have collected several cases of this), which makes their revival more difficult. Among themselves they do not form associations, but combinations. If associations ever occur they are extremely rare and limited in character. Hunger and thirst are special, indecomposable states. Disgust and fatigue are revived easily enough, and, as we have seen, by almost every one; but it must be remembered that these states are composed of somewhat heterogeneous—sensory and motor—elements, and that they approach the character of aggregates. This antithesis between the first two groups and that of odours, tastes, and internal sensations, depends, no doubt, on certain physiological conditions. As we can do nothing but hazard conjectures on this point, it is better to abstain. 2. The second theory stated above—viz., that the revivability is in direct ratio to the motor elements included in the image—is more open to question. I only give it as a _partial_, secondary, subsidiary explanation, applicable to many cases but not to all, and allowing of numerous exceptions. Since we have to do with an empirical law—a pure generalisation from experience—we must test it by facts, in order to fix its bearing and value. This rapid examination will justify my restrictions and reservations. Among all our impressions, those of sight and hearing are those most easily revived. Now, though the visual faculty has at its disposal a very rich, varied, and delicate motor apparatus, this is not the case with the auditory. Considering the superior position of the latter among the senses, it is very poor in motor elements; movements of the head, accommodatory movements of the tympanal membrane, in extreme cases movements of the vocal organs, and, according to the latest hypotheses, a certain function of the semi-circular canals. The difference between the two senses in respect of motor elements is very striking. The sense of smell is much more varied and of greater extent than that of taste. It is far superior as a means of information, yet inferior as regards the sum of the movements at its disposal for exercising itself.