July 8, 1909J 



NA TURE 



41 



The opening chapters deal exclusively with Aus- 

 tralian conditions. Pressure observations are con- 

 sidered first. The mean amplitude of the difference 

 between a number of conspicuous minima and the 

 succeeding maxima in the curves showing the annual 

 variation, amounts to more than seven-hundredths of 

 an inch. When the curves for those stations for which 

 long records are available are compared, they all show 

 a marked similarity, and the important generalisation 

 is arrived at that simultaneous excess or defect of 

 pressure in any one year is a marked feature of the 

 whole Australian continent, and is not restricted to 

 any one particular portion of this area. Coming 

 next to the rainfall observations, an examination of 

 the curves leads to a similar conclusion. Years of 

 low rainfall are, broadly speaking, years of deficiency 

 over the whole continent, and in years of excess the 

 excess is also general. Moreover, a comparison of 

 the rainfall and pressure curves suggests very strongly 

 that periods of high pressure are periods of low- 

 rainfall, and vice versa. These are generalisations of 

 great importance, for they introduce a great simplifi- 

 cation, and correspondingly facilitate the further 

 study of Australian weather conditions. In view of 

 the few data available in proportion to the area 

 considered, a meteorologist, arguing from analogy, 

 might be disposed to regard these as hasty generalisa- 

 tions. The extraordinary variability of rainfall 

 in other parts of the world is well known, and 

 for its adequate study a large inass of information 

 is essential. When the necessary figures are forth- 

 coming we find that even within the narrow limits of 

 our own islands there are very conspicuous differences 

 between the north of .Scotland and the south of 

 England. Australian conditions are, however, dif- 

 ferent. As Dr. Lockyer points out, the weather of 

 the continent is dominated primarily by anticyclones 

 travelling from west to east. In years of high 

 pressure these anticyclones are found to embrace a 

 wider area, and thus the low-pressure systems which 

 skirt their edges and bring rain to the northern 

 districts in summer, and to the southern ones in 

 winter, affect the land area to a smaller extent. 



In discussing these questions of correlation, whether 

 it be between variations of the same element at 

 different places or between different elements. 

 Dr. Lockyer uses the similarity between two curves 

 as his standard of measurement. The points 

 of resemblance to which he directs attention are, 

 indeed, striking. At the same time, the reader 

 feels a desire for a more definite expression of 

 the relation between the elements under comparison. 

 When we come to the correlation between the 

 Australian curves and those for other parts of the 

 world, which takes up much of the later part of the 

 work, this becomes more imperative. Thus, on 

 p. 72, after discussing the striking resemblance 

 between the pressure changes at Adelaide and those 

 of Bombay or Batavia, we read, " While the Cordoba 

 curve is nearly the inverse of Adelaide — the curve 

 for the Cape seems to be intermediate, being more 

 inclined to be similar to the Australian type of varia- 

 tion than that of South xAmerica." The intermediate 

 between two curves which are inverse to one another 

 sliould b.e a straight line. If it is meant that the Cape 

 curve follows now the variations of Adelaide and 

 now those of Cordoba, it becomes a matter of im- 

 portance to have some means of comparing the 

 degrees of similarity in the two cases. 'Superpose 

 any two arbitrarily drawn curves showing fluctuations 

 of approximately the same amplitude, and we are sure 

 to find that some of the maxima and minima agree. 

 Can we say by how much the correlation between 

 the curves we are discussing exceeds that between 

 curves drawn arbitrarily? 



NO. 2071, VOL. 81] 



The question of periodicity naturally comes in for 

 discussion. After eliminating a variation of short 

 period by taking means of groups of four years. Dr. 

 Lockyer claims that the smoothed curves for Australia 

 show a periodicity of nineteen years. It is true that 

 there are conspicuous maxima in 1868 and 1897, and 

 minima separated bv about the same number of years, 

 but this does not of itself prove a recurring periodicity, 

 and the case is not advanced by drawing a 

 " hypothetical " curve through the points of maxi- 

 mum in which an intervening secondary maximum 

 is disregarded and replaced by a principal minimum. 

 The occurrence of a similar interval between . the 

 maxima in the pressure curve for South America, but 

 of other epoch, is suggestive, but the question of the 

 connection between the two continents remains one 

 for further study. 



.\ highly suggestive and interesting chapter on the 

 air movements over the three great land areas of the 

 southern hemisphere points put some interesting 

 similarities between the pressure distribution and the 

 incidence of rainfall of the three continents. The 

 volume also contains an interesting comparison of the 

 flow of the Murray river with the rainfall, and of the 

 frequency of southerly " Bursters " with the varia- 

 tions of pressure. The work thus ranges over a wide 

 field. It offers much that is new, and brings together 

 from a common point of view much information that 

 has hitherto been scattered in a number of individual 

 papers. R. G. K. L. 



POSSIBILITY OF AN EXTRA-NEPTUNIAN 

 PLANET. 



IX/r G.AILLOT has contributed an admirable note 

 -'■'•'•• on this subject to the Couiptes remius (March 

 22). A summary of his calculations is set forth 

 so clearly as to be easy to follow-, and if we have one 

 regret it is that he has not published the discordances 

 between observed and tabular positions that neces- 

 sarily form the basis of his work. We suppose that 

 the Comptes rendus do not admit masses of tabular 

 matter, and we w-ish to express the hope that M. 

 GaiUot will publish this information somehow or 

 other. 



A review recently appeared in Nature (June • 17, 

 p. 463) on Prof. W. H. Pickering's calculations. We 

 there maintained that Prof. Pickering's supposed 

 planet " O " could not possiblv produce sensible per- 

 turbations in Liranus. Now, I\I. Gaillot and Prof. 

 Pickering both locate their hypothetical planets in the 

 same part of the sky. M. Gaillot 's mass is five times 

 that of the earth, or tw-o and a half times that of Prof. 

 Pickering's " O." A reader of the previous review 

 will see that M. Gaillot's planet w-ould, therefore, 

 produce in Uranus inequalities e.xceeding a second of 

 arc. We suspect that Prof. Pickering has made some 

 numerical mistake in estimating the i-i-iass of his 

 planet " O," and, if he can rectifv this, we should 

 then have two independent researches in practical 

 agreement. M. Gaillot's result is, however, suffi- 

 ciently confirmed by the analogy from inner planets 

 developed in the previous review. 



The iniportant question now arises, " Are the 

 observed discordances sufficiently large to point un- 

 mistakably to some unknown planet? " It is clear 

 that an inequality with a coefficient of one second of 

 arc appears to exist in the observations; but the 

 elliptic constants of the orbit of Uranus are arbitrary, 

 the observations are liable to small errors, and the 

 thtorv of the action of known planets is not perfect. 

 All this shows how unsafe it would be to assert the 

 real existence of the inequality which would in its turn 

 demonstrate the existence of an unknown planet. We 



