It follows from this theory that the larger and more convex the eye, the wider will be its field of vision, while the smaller and more numerous are the facets, the more distinct will be the vision (Lubbock). The theory is certainly supported by the shape and size and the immense number of facets of the eye of the dragon-fly, which all concede to see better, and at a longer range, than probably any other insect. Müller’s mosaic theory was generally received, until doubted and criticised by Gottsche (1852), Dor (1861), Plateau, and others. As Lubbock in his excellent summary states, Gottsche’s observation (previously made by Leeuwenhoek) that each separate cornea gives a separate and distinct image, was made on the eye of the blow-fly, which does not possess a true crystalline cone. Plateau’s objection loses its force, since he seems to have had in his mind, as Lubbock states, Gottsche’s, rather than Müller’s, theory. Müller’s theory is supported by Boll, Grenacher, Lubbock, Watase, and especially by Exner, who has given much attention to the subject of the vision of insects, and is the weightiest authority on the subject. Gottsche’s view that each of the facetted eyes makes a distinct image which partially overlaps and is combined with all the images made by the other facets, was shown by Grenacher to be untenable, after repeating Gottsche’s experiments with the eyes of moths, in which the crystalline cones are firm and attached to the cornea. He was thus able to remove the soft parts, and to look through the cones and the cornea. When the microscope was focussed at the inner end of the cone, a spot of light was visible, but no image. As the object-glass was moved forward, the image gradually came into view, and then disappeared again. Here, then, the image is formed in the interior of the cone itself. Exner attempted to make this experiment with the eye of Hydrophilus, but in that insect the crystalline cones always came away from the cornea. “He, however, calculated the focal length, refraction, etc., of the cornea, and concluded that, even if, in spite of the crystalline cone, an image could be formed, it would fall much behind the retinula.” “In these cases, then,” adds Lubbock, “an image is out of the question. Moreover, as the cone tapers to a point, there would, in fact, be no room for an image, which must be received on an appropriate surface. In many insect eyes, indeed, as in those of the cockchafer, the crystalline cone is drawn out into a thread, which expands again before reaching the retinula. Such an arrangement seems fatal to any idea of an image.”