and 94a unquestionably have resemblances in general shape, in the two pedals which are transformed into feet, in vertical lines within the periphery. But why should the word “harp” bring a woman’s skirt and feet peeping beneath it? Perhaps we shall be told it is because a woman plays on a harp. A _woman_ does, yes, but not half a woman, and that half standing so that her skirt takes the form of a harp. If conceivable that “Napoleon” should rouse a vision of an island and induce the drawing of an island, would the island take the shape of half of Napoleon’s body? The mind, conscious or subconscious, does not act in that fashion. Again, the percipient’s drawing which was the sequel to the agent’s balloon (Figs. 95, 95a) is not by itself recognizable as a balloon, and was not recognized by the percipient as a balloon, for she wrote, as we inadvertently neglected earlier to state, “Shines in sunlight, must be metal, a scythe hanging among vines or strings.” The involuntarily whispered word “balloon” would hardly, by any association of ideas, have led to such a reaction; nor would the agent have whispered “half a balloon” or “scythe.” But we _can_ understand how the agent’s eye may have dwelt upon one side or half of the balloon and how his attention may have wandered to the cords, with corresponding telepathic results. See Figures 92, 92a. Here the analogies of form, although imperfect, are nevertheless unmistakable, but what association of ideas could have led from the involuntarily whispered word “chain” or “links,” to “eggs” and “smoke,” or to “curls of something coming out of the end of an egg”? At a later date the agent drew a mule’s head and neck, with breast-strap crossing the lower part of his neck, forming a strip curving very slightly up from the horizontal. The percipient’s drawing is of the head and part of the neck of a cow, turned in the same direction. The long ears of the mule have become the horns of the cow, and matching the breast-strap of the mule there appears a narrow horizontally extended parallelogram in front of the cow’s neck and extremity of its muzzle, which last the percipient seemingly tries to explain by the script “Cow’s head in ‘stock.’” But if the agent involuntarily whispered “mule,” it would hardly suggest a cow, if he whispered “long ears,” it should not have resulted in long horns, if “breast-strap” or “strap” or “harness,” this would hardly bring as its reaction the narrow parallelogram, which, whatever it is, is manifestly no part of a harness. The resemblances in shape are distinct and unmistakable, but they are incomprehensible as the result of overheard whispering. Or look again at Figures 78, 78a. The percipient, especially in the first of her two drawings, very nearly reproduces the original, but the barb of the fishhook has become a tiny flower with a curving stem. The resemblance in shape is exceedingly impressive, but what words could have been