A well-known skilled observer of test objects[41] says: “Practically the resolving power of our achromatic objectives on lined objects reached their maximum in the late Dr. Woodward’s hands. _Amphipleura pellucida_ was then, as now, the finest known regular structure of the diatoms. There appeared then nothing more to be gained in resolution when one of the apochromatic 1/12-inch objectives of Zeiss, with its entire absence of colour, passed into my hands, and I soon became convinced that it possessed the power of separating the different layers of structure in the valve, beyond the grasp of the dry-objective. The result of this increase of power enabled me to split up, as it were, the one plate of silex forming the valve of _Pleurosigma formosum_ into three layers, and which had never before appeared to be possible; proving, in fact, that magnification without corresponding aperture is of little or no account.” “The intimate structure of these test objects,” says Mr. Smith, “is built up on one plan, each being composed of two or more layers, (1) a valve with two layers, as in _Pleurosigma balticum_; (2) two layers with a grating and secondary markings placed diagonally, as in _Pleurosigma formosum_; (3) with two layers of a net-like structure, as in _Pleurosigma angulatum_, the fineness of the striæ or gratings of which measure the 1/50000th of an inch. Five other diatoms afford evidence of this compound structure. The presence of beads or hemispheres in one of the focal planes, and depressions or pits in another, are emphasised in the micro-photograph itself; reduced portions of the valve are represented in Figs. 204 and 204_a_.” A portion of a diatom valve, _Pleurosigma angulatum_, micro-photographed on a higher scale of magnification, 4,500 diameters, is given further on. [Illustration: Fig. 205.--Sections of an old-fashioned Glass Tumbler, from photographs by the late Mr. R. Beck.] _Errors of interpretation_ arise either from the small cones of illumination afforded by the dry-objective, or the oblique illumination formerly resorted to for the resolution of these difficult test objects, and several of the lights and shadows resulting from the refractive power of the object itself. But the most common error is that produced by the reversal of the lights and shadows resulting from the refractive powers of the object itself. To make this clear, I reproduce two reduced photographs of a small section of an old-fashioned glass tumbler, covered externally with numerous hemispheres, illuminated by transmitted light (Fig. 205).