268 



NATURE 



[July 2 1, 1887 



likely to be misleading. This too only represents an arithmetical 

 result, and signifies that as in the series of twenty-two observations 

 there is an average departure of ±5*2 inches from the mean of all 

 the measurements, the probability is that this mean will be 

 within ± o'94 inch of the truth, so far as those measurements 

 are to be trusted. 



For these reasons I am quite unable to follow the arguments 

 by which it is sought to connect the amounts of these two 

 "probable errors,'' or to see how they can in any way indicate 

 "the relative probability of this particular variation being the result 

 of a periodic law, and of its being a mere fortuitous series of 

 variations from a constant average." 



Neither does there appear to be any justification for assuming 

 that the relative probability of the truth of two hypotheses is 

 represented by the inverse ratios of the probable errors of results 

 derived from them. Still less is there any ground for saying that 

 because the particular series of quantities under discussion relates 

 to a period of twenty-twoyears, the relative probability just alluded 

 to is thereby increased to the twenty-second power of that ratio, 

 or from about If to I, to 655 to I. It is no doubt true that if 

 the probability of an event occurring once be represented by the 



fraction -, the probability of its recurring n times in succession 



will be represented by ( — j ; but I fail to see how this affects 



the question at issue. Richard Strachey. 



July II, 1887. 



Is Cold the Cause of Anticyclones .' 



In a review of Loomis's papers in this volume of Nature, 

 p. 2, occur the following sentences: — "While all, or nearly 

 all, of the high pressure of anticyclones may be accounted for 

 by the very low temperatures which overspread the same region 

 at the same time along with the resulting upper currents con- 

 centrating upon them from adjoining cyclonic regions, it is quite 

 different with the low pressures of cyclones. In the case of 

 cyclones the problem is complicated by the strong winds, the 

 copious precipitation, and the ascending currents, which affect 

 the results in ways which no physicist has yet been able to 

 explain." 



This induction of Loomis's, that anticyclones are largely the 

 result of cold, which the reviewer here repeats, is in entire 

 opposition to the deductive views of Ferrel, and I think the 

 discrepancy is to be found in the method used by Loomis in 

 drawing his inductions. In order to investigate the cause of 

 anticyclones, Loomis selected only decided areas of high pres- 

 sure, and as a consequence his data were almost entirely confined 

 to the winter months, when the temperature accompanying anti- 

 cyclones is always low. If, however, he had selected more 

 moderate anticyclones, he would have found that in summer 

 anticyclones in the United States are sometimes accompanied by 

 intense heat (90° F. or more). This is especially so in periods 

 of drought. Under these conditions the approach of a cyclone 

 with rain brings a most refreshing cooling. Furthermore, Hahn, 

 attacking the problem by a different method, has obtained 

 results apparently directly opposed to this induction of Loomis. 

 Hahn made a careful study and cimparison of the observations 

 obtained last autumn and winter on the Sonnblick and at 

 adjacent mountain and valley stations (see Meteorologische 

 Zeitsckrift, February and April 1887) . One of the most marked 

 results found was that at heights exceeding 1000 metres above 

 sea-level there was an increase of temperature during anti- 

 cyclones, while a decreased temperature was only found in 

 valleys and near the general level of the earth's surface. Hahn's 

 average results show that the temperature on the Sonnblick, 

 height 3090 metres, rose from an average of - i6'''4 C. at the 

 average barometric pressure of509'i mm. to - 7°*7 C. at the 

 barometric pressure of 529'3 mm. ; while at the same time the 

 average temperature at Schafberg, height 1776 metres, rose 

 from -9°o C. to + i°-4 C. ; but on the contrary at Zell-a-See, 

 height 754 metres, the average temperature fell from - 5^*9 C. 

 to - 8°'9 C. These results, which show that the larger portion 

 of the atmosphere is warmed instead of cooled within the area 

 of an anticyclone, seems entirely destructive of Loomis's 

 hypothesis that the cooling of the air near the earth's surface is 

 the chief cause of the anticyclone. Hahn's results, however, 

 indicate that the cooling of the air near the earth's surface does 

 increase the pressure somewhat. Thus in October 1886 the 

 barometric minimum occurred at all the stations, both mountain 



and valley, on the 17th ; while the barometric maximum occurred 

 at all of the stations on the 30th. When the difference in 

 pressure between the mountain stations at the time of barometric 

 minimum was compared with the difference in pressure between 

 the same stations at the time of barometric maximum, it was 

 found almost exactly the same ; but the difference in pressure 

 between the valley stations and the mountain stations 

 was about 5 mm. greater at the time of maximum 

 pressure than at the time of minimum pressure. Hahn 

 refers this greater range of pressure at the valley stations 

 to the decreased temperature at valley stations during anti- 

 cyclones, but this still leaves a range of pressure of nearly 

 20 mm. , which the decreased temperature entirely fails to explain. 

 These results of Hahn are in entire accord with the results 

 obtained at Blue Hill Observatory (640 feet above sea), as com- 

 pared with Mount Washington and with stations near sea-level, 

 which indicate that the cooling in anticyclones is almost entirely 

 confined to within a few hundred feet of the earth's surface (see 

 Science, vol. viii. pp. 233 and 281). 



In the light of these facts, it seems more reasonable to assume 

 that the warmth found on mountains and the cold in valleys, 

 accompanying anticyclones, are the result, rather than the cause, 

 of anticyclones. Such researches as those of Langley on the solar 

 heat, and of Hahn on the distribution of temperature pressure, 

 &c., in different planes of the atmosphere, indicate that the 

 effect of the sun's heat on the atmosphere is far more complex 

 than some of our text-books on meteorology would have us 

 believe ; and instead of the cause of anticyclones being as simple 

 as the reviewer of Loomis's article states, it seems probable that 

 we shall understand the phenomenon of the anticyclone only 

 when we master the problem of the cyclone. 



H. Helm Clayton. 



Blue Hill Observatory, Boston, Mass., U.S., June 28. 



Physiological Selection. 



I AM perhaps in a position to contribute something practical 

 to the discussion upon Mr. G. J. Romanes's proposal of physio- 

 logical selection as an improvement upon Darwin's natural 

 selection. 



I failed to meet with Mr. Romanes's paper in the Journal of 

 the Linnean Society, and I confess that I did not gain a very 

 clear idea of what he meant by physiological selection, until I 

 read his article in the Nineteenth Century for January. His 

 main difficulty appears to be the intercrossing with parent 

 stock, which he thinks would prevent the survival of any 

 varieties naturally selected to become species. Now, my 

 Australian bush experience of the habits of animals and birds 

 satisfies me that this difficulty is mainly, if not entirely, imaginary, 

 and that Nature amply provides against the supposed intercross- 

 ing. Any person who has observed the habits of semi- 

 domesticated stock, such as horses and cattle, can scarcely fail to 

 know that migration is a general practice of one sex, and a 

 frequent one of the other. The old always hunt the young males 

 and strangers entirely away to form herds and families of their 

 own, and thus the supposed intercrossing is by one sex at least 

 effectually obviated. But it is also frequently the case that 

 young fillies and heifers, at the same season, take to wandering, 

 for less evident reasons, far from their accustomed haunts, scores 

 of miles, after which they will stop, and attach themselves to 

 another herd and locality as tenaciously as their parents remain 

 in theirs. This of course further tends to prevent intercrossing 

 with parent stocks. 



I cannot but think that Mr. Romanes's anxiety to find a solu- 

 tion of his difficulty has led him into serious mistakes, which 

 vitiate his treatment of the subject. For instance, he says (page 

 59, Nineteenth Century for January), "The hypothesis of 

 physiological selection sets out with an attempted proof of the 

 inadequacy of natural selection, considered as a theory of the origin 

 of species.^' I was out walking yesterday when I read this, and 

 I wrote in the margin, "The theory of natural selection is one, 

 not of the origin of species at all, but of the preservation of , 

 particular varieties." On reaching home, I referred to the 

 "Origin of Species" (4th edition, 1866), and was certainly 1 

 pleased to find that I had adopted Darwin's precise words 1 

 repeated in several places (see pages 71, 91, 123, &c.). ' 

 At page 91 he says : — " Some writers have misapprehended or 

 objected to the term natural selection. Some have imagined that 

 natural selection induces variability ; whereas it implies only the ' 

 preservation of such varieties as occur, and are beneficial under 



