simply through plotting (as in figure 1) the galactic longitude and the distance projected on the galactic plane. No striking lack of symmetry appears in this diagram, except the almost total absence of bright open clusters in the first 90° of galactic longitude. The mean distance of the 70 open clusters along the plane is 5900 parsecs, all individual values (except for the Pleiades) lying between 400 and 16,000 parsecs. By taking the dip of the central line of the Milky Way1 as 1o, and the distance of the sun above the plane as 60 parsecs,17 the distance of the stars and star clouds that enter Newcomb's FIG. 1. DISTRIBUTION OF OPEN CLUSTERS IN THE GALACTIC PLANE = (visual) determination of the position of the galactic circle is of the order of 60/sin 1° 3500 parsecs. Although the uncertainty of this value is large, it seems reasonable to infer that the open clusters are intermingled with the non-cluster stars of the galactic stratum. The diagram in figure 2 illustrates, for all the open clusters, globular clusters, and Cepheid variables falling between galactic longitudes 290° and 360°, the distances projected on the galactic plane plotted as abscissae against the distances from the plane as ordinates. This region of the sky, containing the great Milky Way clouds of Ophiuchus, Sagittarius, and Scorpio, is symmetrical about the point that is indicated by the distribution of globular clusters as the center of the galactic system. It is mainly the absence of globular clusters from low galactic latitudes throughout this interval of 70° in longitude that gives rise to the phenomenon of a region of avoidance. The diagram shows that the distribution of stellar material is probably fairly continuous along the galactic plane; from the local cluster the Cepheid variables (and various other types of highly luminous galactic stars) extend to the nearer star clouds and open clusters, and the latter are recorded among the more distant star clouds along the plane nearly as far as the center of the system of globular clusters. FIG. 2. DISTRIBUTION OF THE GLOBULAR CLUSTERS (ASTERISKS), OPEN CLUSTERS (OPEN Ordinates are distances from the galactic plane; abscissae are distances along the plane If, as appears very probable, the system of globular clusters outlines the galactic system, why do we not find large numbers of open clusters in the vicinity of and beyond the center, between the two halves of the assemblage of globular clusters? Nearer the sun there are some 70 open groups within the mid-galactic segment—the segment which appears to be their natural and only domain. Is the distant central region that is devoid of globular clusters also in part avoided either actually or apparently by open clusters? The observed scarcity of open clusters in this direction leads us to question the reality of the avoidance; it may be that patches of obscuring material conceal both open and globular clusters, as well as many of the more distant stars, perhaps thus playing a large part in the conspicuous rifts in the Milky Way. Barnard's dark markings, recently catalogued,18 do not furnish direct evidence of this obscuration, for, singularly enough, more than half of his objects fall outside the region of avoidance, if we exclude one small region near Messier 11 for which 30 separate positions are listed. It appears that most of the markings may be affiliated with the local cluster, and at no great distance from the sun. Thus in the TaurusOrion region, two-thirds of the dark markings have negative galactic latitudes, lying on the average more than 10° south of the galactic circle; in the opposite region the latitudes are largely positive, the dark markings in Ophiuchus and Scorpio lying intermingled with the seemingly unaffected globular clusters. Along the middle line of the region of avoidance relatively few markings are recorded. Indirectly, however, in Barnard's nebulae we have an argument favoring the hypothesis that globular clusters are concealed in midgalactic regions, for, if a considerable amount of obscuration is associated with the relatively small local cluster, it suggests that such material may also be common in other stellar regions. Although the star counts in typical open and globular clusters fail to reveal as yet the presence of such obscuration, the distribution of stars in the Magellanic Clouds suggests the possibility of its presence there. Another point of considerable weight against a real absence of globular clusters from the region of avoidance is the difficulty and improbability of such a dynamical condition. The distribution of globular clusters in space shows their very close relationship to the Galaxy; the average velocity and probable mass both appear to be very great; the possibility, therefore, of repelling a globular cluster from the stellar stratum, or completely transforming it during a single passage, seems remote. From a gravitational standpoint we should naturally expect the frequency of clusters to be greatest at or near the galactic plane, and that many oscillations must occur before the hypothetical assimilation and transformation is completed for an average globular cluster. Of the several arguments favoring the reality of the empty zone, we recall that the most important are the completeness of the observed absence at all distances from the sun, and the various suggestions of immediate genetic relationship between the external globular clusters and the open clusters within, including the observation that the globular clusters nearest the plane appear to be the most open. 1 Shapley, Harlow, Mt. Wilson Contr., No. 152, 1917 (1-28), pp. 22, 23 and footnotes; No. 157, 1918 (1-26), p. 10; No. 161, 1918 (1-35), sections I, II, and III. 2 Ibid., No. 157, 1918 (1–26), p. 12; No. 161, 1918 (1-35), section IV. 8 Ibid., No. 157, 1918 (1-26); No. 161, 1918 (1-35), sections VII and VIII. Ibid., No. 161, 1918 (1-35), section VIII; No. 157, 1918 (1–26), pp. 12–14. "But whether the absence of clusters is real or only apparent, we must remember that the assignment of a definite thickness of three or four thousand parsecs to the galactic segment is mostly a matter of convenience and approximation; the intention is merely to suggest that practically every known object except spiral nebulae and globular clusters is within a thousand parsecs or so of the central plane of a greatly extended, indefinitely bounded stellar stratum. • Shapley, Harlow, Mt. Wilson Contr., No. 152, 1917 (1–28), fig. 1; No. 161, 1918 (1–35), section V. 1 Ibid., No. 152, 1917 (1–28), p. 22, footnote 2. Ibid., No. 115, 1915 (1–21), p. 11 and fig. 1. 9 Pickering, E. C., Ann. Obs. Harvard Coll., Cambridge, Mass., 26, 1891 (260–286). 10 Adams, W. S., and van Maanen, A., Astr. J. Albany, N. Y., 27, 1913, p. 187. 11 Melotte, P. J., Mem. R. Astr. Soc., London, 60, 1915 (175–186). 12 Kapteyn, J. C., Mt. Wilson Contr., Nos. 82 and 147, Astrophys. J., Chicago, Ill., 40, 1914 (43-126), 47, 1918 (104–133, 146–178, 255–282). 13 Plummer, H. C., Mon. Not. R. Astr. Soc., London, 73, 1913 (174–191). 14 Charlier, C. V. L., Meddelanden Lunds Astr. Obs., Lund, Series 2, No. 14, 1916 (1–108). 15 Bailey, S. I., Ann. Obs. Harvard Coll., Cambridge, Mass., 60, 1908, No. VIII. 16 Newcomb, S., Pub. Carnegie Inst., Washington, D. C., No. 10, 1904 (1–32). 17 Shapley, Harlow, Mt. Wilson Contr., No. 157, 1918 (1–26), p. 23. 18 Barnard, E. E., Astrophys. J., Chicago, Ill., 49, 1919 (1–23). A COMPARISON OF CERTAIN ELECTRICAL PROPERTIES OF ORDINARY AND URANIUM LEAD By P. W. BRIDGMAN JEFFERSON PHYSICAL LABORATORY, HARVARD UNIVERSITY Communicated, June 17, 1919 A comparison of the physical properties of chemical isotopes is of significance because of the light it may throw on the corresponding mechanisms. Comparisons of the properties of ordinary and uranium lead have hitherto been made with respect to the atomic volume,' thermoelectric quality, and emission spectra. No differences have been detected, except possibly a very slight shift in one of the spectrum lines. It is not to be expected that large differences exist with regard to other physical properties, but nevertheless a verification by direct experiment is not without interest. Through the kindness of Prof. T. W. Richards there was made available for me 20 grams of lead of radio-active origin on which he has already published chemical data, and also a similar quantity of puri fied ordinary lead. The radio-active lead was from Australian carnotite, and showed an atomic weight of 206.34, which is therefore 0.41% lower than that of ordinary lead. The theoretical and experimental value for the atomic weight of the pure end product of the disintegration of uranium is 206.08, so that this sample was probably composed of 76% pure isotope and 24% ordinary lead. The radio-active lead contained not over 5 parts in 100,000 of impurity, mostly silver, and the ordinary lead was a trifle less pure, showing also a trace of copper. For the experiments both these samples were formed into wire 0.035 cm. in diameter by cold extrusion through a steel die. The samples were cast into ingots ready for extrusion by Professor Richards, who melted them in hydrogen and then continued the fusion for ten minutes in vacuum. The measurements recorded here are comparisons of the pressure coefficient of electrical resistance, temperature coefficient of resistance, and specific resistance. The comparison of pressure coefficient of electrical resistance was made with more accuracy than the other measurements because a specially adapted apparatus designed for another purpose was available. In order to compare the pressure coefficients, approximately equal lengths of the two varieties of lead were wound non-inductively on either end of a bone core, which was placed in the pressure chamber. The two terminals of each wire were soldered to independent leads which were brought through the walls of the pressure chamber through an insulating plug of a design essentially like that previously described, except that there were three, instead of one, insulated stems through the plug. The two wires were made the two extension coils of a Carey Foster bridge. In this way the difference of the pressure effects on the two coils could be measured. The absolute value of the pressure effect on ordinary lead had been previously determined with sufficient accuracy. The apparatus for producing pressure was the same as that previously described.5 Readings were made to 12,000 kgm./cm.2 at 1000 kgm. intervals, with increasing and decreasing pressure, and at two temperatures, 25° and 85°. An independent set of readings was made to determine the effect of pressure on the resistance of the leads, which turned out to be almost negligible. At 25° the decreases of resistance of the two kinds of lead under 12,000 kgm. were the same within 0.02% of the total decrease, and at 85° within 0.03%. Assuming that the possible error in reading the slider settings of the Carey Foster bridge was 0.1 mm. and that the errors conspired in the most unfavorable way, the possible |