VOL. 19, 1933 : H. SHAPLEY 29 where irregular and bizarre ganglionic masses exist in the complete ab- sence of myotomes. The reduction in the number of ganglia and spinal nerves following somite removal, and especially an increase following the interpolation of an additional somite, constitute strong experimental support of Leh- mann's view that the normal segmentation of spinal ganglia and nerves is determined by the mesodermal metamerism. There is no support of the generally held view that nervous metamerism is intrinsic, at least in the spinal cords of the forms which we have studied. 1 Lehmann, F., Jour. Comp. Neur., 49 (1927). 2 Detwiler, S. R., Jour. Exp. Zool., 54 (1932). 3 Detwiler, S. R., op. cit. Lewis, W., Anat. Rec., 4 (1910). 5 The normal brachial nerves constitute the third, fourth and fifth spinal nerves. 6 Lehmann, F., op. cit.

A CONTRIBUTION TO THE STUDY OF GALACTIC DIMENSIONS By HARLOW SHAPLEY HARVARD COLLEGE OBSERVATORY, CAMBRIDGE, MASS. Read before the Academy, Monday, November 14, 1932 1. In our study of the dimensions of the galactic system, as well as in our exploration of the metagalaxy, we have been gradually driven out of the and low galactic latitudes by increasing knowledge of the structural complexities along the Milky Way, and especially by the confusion and uncertainty caused by general and selective absorption of light near the galactic plane. Many years ago three factors had already pointed to the troubles in store for attempted analyses of galactic di- mensions along the galactic circle. They were Barnard's and Wolf's work on the distribution and nature of the dark nebulae, the general consideration by many astronomers of the great rift in the -to- part of the Milky Way, and the recognition of the "region of avoidance," the last of special importance in showing that star clusters of any kind in low latitudes can give, when their magnitudes are appro- priately corrected for light absorption, an indication of only minimum dimensions of the and the minimum distance to the center. The region of avoidance is the mid-galactic belt, nearly ten degrees in width, in which practically no globular star clusters are found. There is another well-known region of avoidance, wider and more remote and possibly quite distinct, that affects the apparent distribution of the external in low galactic latitudes. There can be little doubt that con- Downloaded by guest on September 27, 2021 30 ASTRONOMY: H. SHAPLEY PROC. N. A. S. siderable numbers of clusters are behind the nearer obscuring material, mostly in Sagittarius, Scorpio and . 2. I have pointed out on various occasions that certain globular clusters are partially obscured.' The dark nebulosity is recognizable in the vicinity of some of these affected clusters, and the stars in others seem to be abnormally faint for the angular dimensions of the systems. Presumably the light is scattered as well as obstructed and color excesses should appear, though nothing peculiar is found in the colors in Messier 22 (galactic latitude -9°) from an extended survey of the magnitudes of individual stars.2 Recently Stebbins has found, from colors determined with a photo- electric photometer and standard screens, that globular clusters in low latitude are systematically redder than those away from the galactic circle.3 Probably this reddening is a true color excess and if so is further evidence of the light scattering in low galactic latitudes, already shown by various investigators of distant galactic clusters and Milky Way stars. It is to be noted, however, that Miss 'Cannon's classification of the inte- grated spectra of globular clusters shows later types in low latitudes. Thus we have data from Appendix A of