D O C U M E N T 2 7 0 A P R I L 1 9 2 6 2 8 5 Interference path-lengths without grating G: Interferences are maintained up to about 50 cm = .[3] Interferences with grid grating (100 divisions/cm), (very accurately divided): In- terferences stay clear to about 7 cm, their sharpness decreases to a minimum at 15 cm, increases again beyond 25 cm, is clear at 30 cm, decreases again around 35 cm, & is very unclear again at 45 cm. A complete vanishing of these interfer- ences is not observable nevertheless, I believe I am seeing a secondary effect here that is connected with grating diffraction. In the next experiments, I am going to try to estimate the intensity of the inter- ference sharpness, in order then to obtain the dependence of the interference on the path difference. I attempted this same experiment with a grid of 200 divisions/cm there I could not observe any secure reduction in the interference sharpness, which I regard as confirmation that the interferences are weakly sustained also for the above grid as a consequence of grating diffraction. Further, I tried out the same experiment with a blackened grating (dark rulings 0.15 mm wide, bright rulings 0.12 mm), with very blurry edges along the dark rul- ings. Here, too, the interferences did not vanish, probably as a result of the imper- fection of the grating rulings. Perhaps it would be good to try the experiment with a grid of 50 divisions/cm. As I cannot use the full length of the canal ray beam (the interference capacity diminishes toward its end), I was unable to try out your last suggestion in this grat- ing trial.[4] In the tilted-mirror experiment, I would like to pose the question of whether in- terferences of equal inclination (telescope at ∞) are required. I imagine the proce- dure as follows: At a determined path difference that is as large as possible (e.g., 20 cm), the in- terferometer is set to clear interferences of equal inclination by means of a light source at rest (Hg-light generated by glow discharge). Then one takes the canal ray light source & observes whether the location of clearest interference has displaced as a function of the mirror tilt. The mirrors then are no longer //, as would corre- spond to interferences of equal inclination with a light source at rest instead, they are shifted by the ∠β, whereas the location of the interferences at stays the same.— Can the tilt of the mirrors here be substituted by a tilt in the interferome- ter’s distributor plate (Ann. p. 5. P)? If the experiment is done with a lens of short focal length, as indicated in your last suggestion,[5] then narrow slit diaphragms must necessarily be used in order to suppress the Doppler components. 2( a1 a2)
Previous Page Next Page