May 1, 1888.] 



♦ KNO\VLEDGE ♦ 



151 



if under special atmosplieric conditions there were mists 

 over the waters of the river, or the river being frozen 

 remained snow-covered in spring after the snows liad melted 

 from its shores, the river would be bright on a darker back- 

 ground, and then the difTraction image of the river would 

 appear as a broad band of light between a pair of parallel 

 dark streaks. At other times the river would not be seen 

 at all, the continent with its rivers being more or less 

 enshrouded in cloud and mist. 



It seems reasonable to suppose that (1) during the winter 

 months of the northern hemi.'^phere the rivers on Mars 

 would not be visible, or would at least not be conspicuous ; 

 (i) after the vernal equinox the clouds and mists hiding the 

 continents and rivers in great part from view would melt 

 awav, but would linger longest over the river beds ; and 

 (.3) as summer approached the mists would melt away during 

 the midday hours over the rivers also. The observations of 

 Schiaparelli, as I have endeavoured to interpret them, 

 correspond with this sequence; for (1) in the winter of 

 north Mars no dark streaks are seen, or but few, and those 

 indistinctly; (2) in the spring are seen the parallel dark 

 streaks ; and (3) towards summer the duplicated streaks 

 become single. 



Now during May, June, and July, ]\lars will be passing 

 through the late summer and autumn of his northern hemi- 

 sphere, and it will be interesting to inquire whether the 

 dark single streaks, indicating clear skies over the river 

 beds of Mars, change again into the dark double streaks 

 indicating mist along the river tracks, before passing as 

 winter advances, under those envelopes of mist and cloud, 

 which will hide the rivers altogether from our view. 



I give in fig. 2 the construction for determining the posi- 

 tion of the polar axis of Mars, and the opening of the 

 equator on May 1.* From that date onwards the planet's 

 presentation will not differ importantly while the planet is 

 favourably situated for observation, so that fig. 3, the pro- 

 jection for the inverted telescopic aspect of the planet (on 

 the meridian) will serve well enough to the end of this 

 Martian season. 



WEIGHING THE EARTH. 



(Concluded J'rom. pcuje 12-3.) 



yS-O jijiiiSSj yi HE principle of the method is Olustrated in 

 fig. 1. Here a and h are two small globes 

 at the ends of a uniform rod r r' , suspended 

 in a horizontal position by the cord or wire 

 c C, attached to its centre C. The horizon- 

 tal rod, left to itself, tends to a mean posi- 

 tion which may be called the position of rest, 

 though, as a matter of fact, when the suspension is as 

 delicate as it has to be in the experiments considered, the 

 rod never is at rest, but oscillates constantly, and very 

 slowly, through short arcs on either side of its mean position. 

 A and B are two heavy globes, which can be brought readily 

 into such positions as are shown in fig. 1, where their 

 attractions tend to draw a and h towards them in the direc- 

 tions shown by the arrows. The result of these disturbing 

 influences is to sway the rod r ?•' from what had been its 

 position of rest, when undisturbed, to some new position of 

 rest, as n n' or m m', about which it oscillates as before. 



The processes of observation are as follows : First, the 

 time of oscillation of the undisturbed rod Ls noted, to 

 ascertain the force of torsion which has to be overcome to 

 produce a given displacement. Then, the large globes being 

 brought up to such positions as A and B, their distance 



* See Knowledge for February 1, 1S84 (vol. v.), p. 71, for an 

 account of the simple process to be followed in all such cases. 



from the positions of rest of a and b when these were 

 undisturbed is carefully measured, and the new position of 

 re.st taken up by « and b is ascertained, the times of 

 oscillation being noted throughout, so that any change in 

 the torsion may be recognised and taken into account. This 

 having been done, the globes A and B are removed to the 

 mean positions M, M', and the balls a and b are allowed to 

 return to their position of rest. Then the globes A and B 

 are carried round in the same direction until A is close by b 

 on the left, and B close by a on the right, when their 

 attractions tend to displace a and b in directions contrary to 

 those shown by the arrows : the position of rest of the rod 

 r >■' is next ascertained (the times of observation being 

 throughout carefully noted) as before. From these observa- 

 tions, the attractions of the globes A and B on the balls a 

 and b (or b and a) can be determined, since the times of 

 oscillation indicate the torsion, and the position of rest 

 determines how much of the torsion is overcome by the 

 globes' attraction. By calculating next what the attraction 

 of either globe would be if, instead of being at its measured 

 distance from the neighbouring ball, it were at the earth's 

 centre, and comparing this with the known attraction of the 

 whole earth, we can ascertain how much the whole mass of 

 the earth exceeds the known mass of the leaden globes : in 

 other words, we can ascertain the mass of the earth, and 



Fig. 



.,.-•' M' 



1. — Illustrating the Principle of the Cavendish 

 Experiment. (Baili/.) 



therefore — its volume being known — we can determine its 

 mean density. 



Cavendish, near the end of last century (" Phil. Trans.," 

 1 798). applied this method in a form somewhat simpler than 

 that described, but depending practically on the same prin- 

 ciples, deducing a mean terrestrial density of 548. Hutton, 

 re-examining C'avendi.sh's experiments, reduced the deduced 

 density to 5 32 ; but many prefer Cavendish's own treatment 

 of his observations. 



The experiments made by Cavendish were not very 

 numerous, neither were those of Reich, of Freiberg, who in 

 1838 deduced by the .same method a mean terrestrial density 

 of 5 '44. We must attach much more weight to the ex- 

 periments made by Francis Baily in 1838-42, since they 

 were not only conducted with singular care and caution, 

 but were obtained in many diflerent ways (though, of course, 

 all by the same general method) and by a ver}- large number 

 of experiments. It will be well to give a few details respect- 

 ing Baily's work, that the trustworthiness of his estimate 

 of the earth's weight may be fully recognised. 



The rod rr', fig. 1, of 6.'f feet long, was of light wood (in 

 nearly all the experiments), and was suspended in various 

 ways in the different experiments, viz., on single copper wire 

 •0178 inch and '0219 inch in diameter; on two parallel iron 

 wires 0'177 inch, 0367 inch, and 0'415 inch apart; on two 

 brass wires 0-380 inch and 0'415 inch apart ; and on two 

 silk fibres 0177 inch, 0-307 inch, 0380 inch, and 0-415 inch 

 apart. At the ends of the rod were attached balls of dif- 



