November 4, 1909J 



A'.-4 TURE 



13 



streams, and thus a verification was obtained of the 

 result that the two streams showed no difference as 

 regards the magnitude or type of spectrum of the 

 stars in them. Of 1800 stars examined, iioo belonged 

 to the first stream, 600 to the second, and the remain- 

 ing 100, which could not be assigned to either, showed 

 no motion of a systematic character. The large pro- 

 portion of stars belonging to the first stream arises 

 from the mode of selection according to the magni- 

 tude of proper motion. Kapteyn's and Eddington's 

 result, that when stars are taken without selection 

 they are equally divided between the two streams, is 

 used to determine the ratio of the stream velocities. 

 When this is determined the apparent movement in 

 two streams, as seen from the earth, is replaced by the 

 solar motion and two streams moving in opposite 

 directions relative to their centre of gravity. 



There is at first sight considerable difference between 

 Kapteyn's description (followed by Eddington and 

 Dvson) of the systematic movements of the stars, and 

 tliat of Schwarzschild. The dual character of 

 Kapteyn's system should not be unduly emphasised. 

 Division of thg stars into two groups was incidental 

 to the analysis employed, but the essential result is the 

 increase of the peculiar velocities of stars towards one 

 special direction and its opposite. It is this same 

 feature, and not the spheroidal character of the distri- 

 bution, which is the essential of Schwarzschild's 

 representation. The results obtained by the two 

 methods agree very closely. Defining the "apex" as 

 the direction of the sun's motion relative to the centre 

 of gravity of the stars, and the "vertex" as the direc- 

 tion of motion of one stream relatively to the other 

 (Kapteyn) or the major axis of Schwarzschild's 

 spheroid, the accordance of the different results is 

 shown in the following table : — 



Apex Vertex 



R A. Dec. R.A. Dec. 



Kapteyn — Bradley stars — .., 91 + 13 



Eddington — Groombridge stars ... 266 + 31 ... 95+ 3 



Schwarz^child—Groombridge stars ... 266 + 33 ••• 93+ ^ 



Dyson — Stars of large proper molior. 281+42 ... 8S + 24 



Keljawsky— Porter's stars 281 + 36 ... S6^24 



Eddington — Zodiacal stars — ... IC9+ 6 



It may be noticed that the Groombridge stars gave 

 almost identical results by the methods of Eddington 

 and Schwarzschild, and that Beljawsky and Dyson, 

 whose material was very similar, obtained results in 

 close accord. 



.Although attention may be directed to Kapteyn's 

 observation that the vertex lies in the plane of the 

 Milky Way, it is too soon to offer any explanation of 

 tliese remarkable movements of the stars. To have 

 disentangled them from the irregular proper motions 

 of the stars is itself a very important step. By 

 clearing up the difficulty in the anomalous results 

 previously found for the direction of the solar motion, 

 and by the discovery of systematic movements in 

 which all the stars share. Prof. Kapteyn has made the 

 most important contribution to this branch of 

 astronomy since the lime of Herschel. 



F. W. Dyson. 



THE SEA OF ARAL. 

 T) ECENT explorations in Central Asia, and the 

 -'■^ evidence accumulating from many quarters of 

 general desiccation of that area within historic times, 

 give special interest and value to anything in the 

 >hape of observations of even approximate precision 

 which point towards an opposite conclusion, or to a 

 conclusion that variations in the amount of precipita- 

 tion, where they occur, are more or less local and 

 NO. 2088, VOL. 82] 



constitute merely a phase which, although it may be 

 of relatively long period, does not represent continu- 

 ous progressive change. The work carried on by 

 L. S. Berg in the Sea of Aral between the years igoo- 

 1906 form an important contribution to the subject of 

 limnology generally, and more particularly to this 

 question of desiccation. The original report on these 

 investigations (Berg, "The Sea of .Aral," St. Peters- 

 burg, 1908) is published in Russian, but students un- 

 familiar with that tongue may acquaint themselves 

 with the present state of knowledge concerning the 

 whole region by means of an article by Prof. Woei- 

 kow, published in the .April number of Petermann's 

 Miitcihtngcn. Prof, ^^"oeikow deals primarily with 

 Berg's observations, and his maps are reproduced, 

 but he uses information derived from other sources, 

 for purposes of comparison. 



The Sea of Aral is situated at an elevation of 50 

 metres above mean sea level, and its area of 63,270 

 square kilometres places it fourth in size amongst the 

 inland lakes of the world. The mean depth is 16 

 metres, and the maximum 68 metres, depths exceed- 

 ing 30 metres occurring only in one small depression 

 in the west and two still smaller ones in the north 

 of the basin. The volume of water is computed to 

 be 1012 cubic kilometres, only slightly greater than 

 that of Lake Ladoga, which has about one-quarter of 

 the superficial area, and about one-tenth of Lake 

 Baikal, which is little more than half the size. The 

 supply of water comes wholly from the two rivers 

 .Amu and Syr, which together deliver, on the average, 

 some 1500 cubic metres per second. Most of the 

 water is derived from the melting of mountain snows, 

 the months of maximum flow being June, July, and 

 August. Berg gives the mean salinity as 1075 /"''' 

 mille : compared with analyses made during the 

 'seventies of last century, which yielded an average 

 of over 12 pro mille, this shows a marked freshening, 

 due, as appears, to an increased volume of water. 



The survey of the Sea of .Aral by .Admiral Butakow 

 in the late 'forties formed the first foundation of 

 accurate knowledge, and there is evidence to show 

 that at the time of that survey the level of the water 

 was relatively high. Few precise measurements were 

 made for a long time afterwards, but it seems certain 

 that after the 'forties a period of falling level began, 

 and continued for some thirty to thirty-five years. 

 Borczow reported diminishing area in 1857. The 

 period 1859 to 1874 is blank, or nearly so. Sewertzow, 

 Subow, and Kaulbars (1873-4) found great shrinkage 

 on comparison with Butakow 's survey, and further 

 comparison with the records of Meyendorff and 

 Ewersman (1820) seemed to justify the conclusion that 

 a general desiccation of this part of Central Asia was 

 taking place continuously. K. Schulz, surveying in 

 the north-eastern end of the sea in 1880, found still 

 further shrinkage since 1874. 



From 1S80 to 1899, when Berg first visited the 

 region, another blank occurs; but in 1S99 Berg found 

 a rise of level in full progress, the height already 

 attained exceeding not only that of 1874 and 1880, but 

 that of Butakow's records in the 'forties. Islands, 

 for example, which appeared in Butakow's map, 

 and which had become peninsulas in the 'seventies 

 and 'eighties, were submerged. Working from the 

 levels of Tillo at Karatmak, Berg estimated a height 

 of i'2i metres in 1901 above that in 1874. Glukhow- 

 skoy found a fall of 71 centimetres between 1874 and 

 1S80, giving a rise from 18S0 to 1901 of about 

 2 metres, or 9 centimetres a year. The rise has con- 

 tinued, and Berg now gives it as 275 metres in 1903, 

 and 3 metres in 1908. The depth of the lake being 

 mostly shallow, this rise corresponds to a very con- 

 siderable increase in area ; the increment in volume 



