OCTOBEB 23, 1908] 



SCIENCE 



575 



tied together and to the mainland by a complex 

 system of tombolos. Some of the drumlins show 

 sea cliffs now abandoned by the waves. From the 

 relations of these cliffs and the more ancient of 

 the beaches, the initial drumlins have been recon- 

 structed. The effect of marine action in cliffing 

 the drumlins and stringing out the eroded ma- 

 terial in successive tombolos has been followed 

 through step by step, until the conditions of 

 to-day were reached. 



The study shows that Nantasket Beach is not 

 the result of the accidental tying together of a 

 few islands without system, but that it represents 

 one stage in a long aeries of evolutionary changes, 

 which have occurred in orderly sequence and in 

 accordance with definite physiographic laws. 



The Acid Extreme of the Oortlandt Series near 

 Peekskill, N. Y.: Charles P. Beekbt, Colum- 

 bia University. 



The rocks of the Cortlandt series are known, 

 through the work of the late Professors J. D. 

 Dana and H. S. Williams. They occupy an area 

 on the Hudson River just south of Peekskill, N. Y., 

 and include a very wide range of granitoid me- 

 dium to basic types of igneous rocks. 



It seems certain that they represent a case of 

 magmatic diiferentiation that includes not only 

 the Cortlandt series, as outlined by Dana and 

 Williams, but also two or three occurrences of 

 typical granite. The granite area borders basic 

 varieties on the northeast side. Actual contacts 

 of the larger masses are not to be seen, but an 

 occasional dike of granite cuts the adjacent diorite 

 and gabbros, indicating a relationship as one of 

 the latest developments. Furthermore, the gran- 

 ite shows consanguinity by its heavy soda content, 

 soda-lime, feldspar predominating. It is, however, 

 a very acid granite and introduces a considerably 

 greater range of rock variety than formerly cred- 

 ited to the Cortlandt series, becoming its acid 

 extreme. 



The Evolution of Bogoslof Volcano in Bering Sea: 

 T. A. Jaggab, Je., Massachusetts Institute of 

 Technology. 



The island consists of four prominent peaks, 

 old Bogoslof at the south, McCulloch Peak steam- 

 ing actively in the middle, Metcalf Cone (some- 

 times called Perry Peak) adjacent to McCulloch 

 in the north, and New Bogoslof or Fire Island 

 ( " Grewingk " ) , a flat table rock at the northwest 

 end of the group. These are now all connected by 

 continuous gravel and sand strips, where in one 

 place there was a broad channel and seven fath- 

 oms of water a year ago. 



McCulloch Peak and Metcalf Cone are both 

 products of the slow pushing up from beneath the 

 waves of a mass of refractory lava, semi-solid, 

 crusting and breaking into blocks as it rises, with 

 only the central portions retaining a semblance 

 of fluidity. 



A series of sketches were shown illustrating 

 the remarkable differences in outline of this island 

 at different intervals from 1826 to 1907. 



In 1796 Old Bogoslof rose. In 1884 New Bogos- 

 lof, Fire Island, came into being and the waves 

 joined the two with bars. In 1891 New Bogoslof 

 was still steaming. In 1906 Metcalf Cone was 

 reported midway between Old and New Bogoslof. 

 In July, 1907, Metcalf Cone had broken in two, 

 and the breaches between the islands were again 

 connected with continuous land. On September 1, 

 1907, McCulloch Peak exploded and was wholly 

 destroyed. 



No such extraordinary story of growth and 

 alteration of an island in the sea has ever been 

 told before, and the changes of the later stages 

 are unique in the annals of volcanology. 



This paper is printed in full in the report of 

 the expedition to Bogoslof. 



Some Curves illustrating Coincident Volcanic, 

 Seismic and Solar Phenomena: Ellswoeth 

 HtTNTiNGTON, Yale University. 

 In discussions of the possibility of some rela- 

 tionship between sunspots and earthquakes or 

 volcanoes, attention has usually been concentrated 

 upon sunspot maxima. Jenson, an Australian, 

 however, has plotted the most important earth- 

 quakes and volcanic eruptions for the last century 

 and more, and on comparing his data with the 

 sunspot curve for the same period finds that there 

 seems to be a grouping of the terrestrial phe- 

 nomena at or near the time of sunspot minima. 

 In order to test the validity of his conclusions 

 another set of data as to earthquakes and vol- 

 canoes, prepared by Mr. R. W. Sayles for quite a 

 different purpose, have been taken and similarly 

 compared with the simspot curve. In this case, as 

 in the other, the grouping of terrestrial phe- 

 nomena at times of sunspot minima is evident. 

 In order to get rid of the personal equation, which 

 enters so largely into such studies, and in order 

 to get rid of temporary or local irregularities, ^11 

 the data of both Sayles and Jensen have been 

 averaged together. By repeated averaging of re- 

 sults as to the frequency and intensity of both 

 earthquakes and volcanoes, the whole body of 

 facts given by the two investigators, for a period 

 of 117 years in one case, and 147 in the other, ha.s 



