October 13, 1910] 



NATURE 



471 



value of the coefficient -0144 + 0076. Thus high pressuii- 

 at Zikawei is accompanied by low temperature, not only 

 there, but also at Nagasaki. Again, taking pressure at 

 Zikawei and rainfall at Keelung, in northern Formosa, 

 450 miles distant, he finds for the value of the coefficient 

 0-925 + 0037. Thus when pressure is high at Zikawei in 

 winter rainy weather in northern Formosa may be con- 

 fidently predicted. The paper includes diagrams, in which 

 the relationships can be distinguished, but the definite 

 numerical index furnished by the correlation supplies the 

 meteorologist with much more satisfactory means of com- 

 paring the similarity with that for other places and of 

 using the results in more extended researches. 



Interest in the diurnal variations of the meteorological 

 elements has been stimulated of late years by the addition 

 to our knowledge of the analysed results of wind velocity 

 for certain representative places, and by the difficulty of 

 bringing into dynamical relation with each other the 

 results for pressure, temperature, and wind. In Bulletin 

 4, 19 10, of the Central Meteorological Observatory of 

 Japan, Y. Tsuiji discusses the daily variation of wind 

 and the displacement of the air at Nagasaki. .After 

 analysing the variation into its harmonic components, he 

 attempts to connect it with the variation of pressure 

 gradient deduced from the mean hourly values of pressure 

 at the three stations Nagasaki, Fukuoka, and Kumamoto. 

 He concludes that the phases of the variation of the com- 

 ponents of wind are almost identical with those of the 

 variation of the corresponding components of pressure 

 gradient. Apart from any effect due to the earth's rota- 

 tion, it would be expected that these phases would difTer 

 by 90", and although the nearness of Nagasaki to latitude 

 30" might account for the identity in the phases for the 

 diurnal wave, it is unlikely that the effect in the semi- 

 diurnal wave can be attributed to such a cause. The 

 author does not attempt to explain the anomaly. .An 

 examination and recalculation of some of the results shows 

 that time is to be measured from i a.m. in the expressions 

 given for the harmonic components, although it is stated 

 that it is measured from midnight. Thus the diagram 

 (Fig. s) showing the diurnal and semi-diurnal waves is 

 seen to be wrong by a comparison of the sum of the 

 deviations at noon taken from the diagram with the aciual 

 values given in Table 1. The fact that the phases of the 

 diurnal term in the south and east components of velocity 

 are almost exactly equal, while those for the semi-diurnal 

 wave differ by 40° only, shows that convection is the 

 predominating factor in determining the diurnal variation, 

 and exercises a considerable influence on the semi-diurnal 

 wave. 



In a paper which appeared in vol. xxxii. of the Annalcn 

 der PItysik Dr. M. Knudsen, of the University of Copen- 

 hagen, describes an absolute manometer for the measure- 

 ment of gas pressures not greater than a few thousandths 

 of a millimetre of mercury. It depends on the measure- 

 ment of the force of repulsion between two plates at 

 different temperatures immersed in the gas, when their 

 distance apart is extremely small compared with the mean 

 free path of the molecules of the gas. This repulsion the 

 author shows is equal in dynes per square centimetre to 

 half the product of the pressure of the gas by the excess 

 of the square root of the ratio of the two absolute tempera- 

 tures of the plates over unity. The instrument consists 

 of a polished copper plate, which is supported with its 

 surfaces vertical by a thin' platinum wire. Opposite part 

 of one surface is the polished end of a fixed copper 

 cj'linder, the temperature of which can be raised above 

 NO. 2137, VOL. 84] 



that of the suspended plate and the case of the instru- 

 ment by sending an electric current through a platinum 

 wire wound on it. The rotation of the moving plate 

 about its suspension is determined by means of a mirror 

 attached to the plate, and thermoineters give the tempera- 

 tures of plate and cylinder. 



Prof. P.\terno, of the University of Rome, has given 

 an account in the Rcviie scicntifiijue of .August 20 of some 

 recent work on colloidal solutions, arising out of an early 

 observation that tannic acid, which behaves as a colloid 

 and produces no osmotic or cryoscopic effects in water, 

 behaves in the normal way as a crystalloid when dissolved 

 in acetic acid. The aqueous solution is a deep brown, 

 although the tannic acid separating from it is white, and 

 the solution in acetic acid is only yellowish. Observations 

 by the ebullioscopic method indicated that the Grignard 

 reagent, magnesium ethiodide, probably exists as a colloid 

 in its ethereal solutions. Several alkaloids were found to 

 dissolve in ether without raising its boiling point, whilst 

 the alcoholic solutions showed a normal elevation. 



In an article in the American Architect for September 28 

 Prof. William H. Goodyear analyses the report of the 

 Pisa Commission on the Leaning Tower, and brings for- 

 ward evidence to show that the figures of the commission 

 are in error. Prof. Goodyear maintains that the tower has 

 mo\'ed 30 centimetres more than the commission has sup- 

 posed, and that this movement had taken place before the 

 date of De Fleury's " Monuments de Pise " in 1859. The 

 actual movement, he says, has been one of 50 cm., not 

 of 20 ; but this movement, which was probably caused by 

 the earthquake of 1846, certainly occurred before 1859. 

 Prof. Goodyear also challenges other results published by 

 the commission. Thus the commission's supposed rate of 

 inclination per metre in 1829 is 865 mm. ; the article gives 

 the real average rate as 84 mm. The supposed present 

 rate of inclination per metre is announced by the com- 

 mission as 92 mm. ; here it is said to be really 945 mm. 

 in 1859. The commission announces 20 cm. additional 

 inclination between 1829 and 1910. There was really, it 

 is said, 50 cm. additional inclination before 1859. The 

 commission announces the rate of increase per metre 

 between 1829 and 1910 as being 5J mm. Therefore, says 

 Prof. Goodyear, if the supposed increase of 20 cm. be 

 divided by the rate of increase per metre, we ought to 

 obtain the axial height on which the increase was figured, 

 viz. the axial height of the tower. " But when this 

 division is made, the resulting axial height is only 

 36-36 m., which is the axial height, according to Cresy 

 and Taylor, of the si-x stories above the first, thus leaving 

 the entire lower story, which is nearly one-fourth of the 

 axial height, out of the computation." 



-An illustrated article on the Portsmouth water works 

 appears in Engineering for October 7. The borough of 

 Portsmouth has recently put into commission a new 

 system of filter-beds and covered service reservoirs. The 

 new works are situated at Farlington, on the side of 

 Portsdown Hill, distant five miles from Portsmouth. The 

 water supply comes from chalk springs at Havant and 

 Bedhampton, where pumping-stations are situated. The 

 water is delivered through three rising mains to the filter- 

 beds, from which, after passing through the service 

 reservoirs, it gravitates to the town. Previous to the con- 

 struction of the new works, the water was delivered from 

 the pumping-stations to two open service reservoirs on 

 Portsdown Hill ; these are now covered, and form part 

 of the reconstructed scheme. It was the practice under 



