tion; and the reservoir may be dismissed with the remark that it is a bottle of stone or glass, 9.7 inches high. The Receiver is a copper circular funnel, 5 inches in diameter, 4 8 inches deep, and terminating in a tube 8.5 inches long and 3 inch in internal diameter. Outside this, and soldered to the bottom of the funnel, is a cylindrical phlange, 2.25 inches deep, and having between it and the tube a space for the reception of the head of the reservoir, which it exactly fits, so that when united a horizontal section through the phlange would disclose three tightly-fitting concentric tubes. The phlange keeps the receiver steady, prevents the rain which falls on the outside of the funnel from leaking into the bottle, and reduces to a minimum the evaporation of the contents of the reservoir. When fitted together the height of the instrument is 14.1 inches; but when in use it is placed firmly in the ground, and should have its top 9 inches above the surface.

Supposed Influence of the Moon on the Rainfall.—That the moon is very influential in, or at least closely connected with, all changes of the weather, is a belief at once widely spread and deeply rooted. Our satellite can neither be full, nor new, nor "fill her horns," without, as is popularly believed, causing or indicating some alteration in the state of the weather. If she is caught "lying on her back," or, in other words, if, when she is less than a semicircle, her cusps are pointed upwards so that the straight line joining them is more or less approximately parallel to the horizontal plane, the fact is supposed to be an indication if not the cause of rain. If she submits to be "towed by one star and chased by another," that is, if she is between and near two conspicuous stars, so that the three bodies are at least nearly in a straight line, the fishermen expect a storm.

Though meteorologists show no favour to these and many similar beliefs, some of them admit that it is neither unphilosophical nor contrary to fact to regard the moon as a meteorological agent. Thus, Sir John Herschel, from his own observations, regards it as a meteorological fact that the clouds have a tendency to disappear under the full moon, and adds that a slight preponderance in respect of quantity of rain near the new moon over that which falls near the full, would be a natural and necessary consequence of a preponderance of a cloudless sky about the full. M. Arago, who concurs in this opinion, states that the expression "the moon eats the clouds," is common in France among country people, and especially among sailors.† The latter philosopher adds that the results obtained from meteorological observations in Germany and in Paris, were that the maximum number of rainy days occurred between the first quarter and full moon, and the minimum between the last * 'Outlines of Astronomy,' par. 432, and note, p. 285. 5th edit. 1858. 'Popular Astronomy,' Smyth's Translation, vol. ii., ch. xxiii., pp. 311-313.

1858.

VOL. VII.

2 K

quarter and new moon; the ratios being 100: 121 4 in Germany, and 100 126 in Paris; but that in the south of France the minimum number of rainy days occurred between the full moon and the last quarter. He concludes with the remark that "the question requires to be examined afresh."*

Having by me an unbroken series of carefully-made rainfall observations from the beginning of 1864 to the present time, I have tabulated the results below so as to show the amount of rain, the number of wet days, and the wet-day rate of rain in each of the four quarters of the seventy-four complete lunations, beginning with the new moon on January 9th, 1864, and ending with January 1st, 1870-a period of 2185 days.

The word "quarter," as used here, may be defined thus:-The first quarter begins with the day of the new moon, and ends with the day immediately preceding that on which, according to the almanac, the moon reaches the first quarter, and so on for the others.

From the foregoing Table, it is obvious that with regard to the three pluvial elements, in South Devon, during the six years ending with January 1st, 1870, the four quarters of the seventy-four moons may be arranged, in descending order, as below:

*Popular Astronomy,' Smyth's Translation, vol. ii., ch. xxxv., pp. 317, 318. + The rainfall of one "second quarter" was exactly 25 per cent. of that of the lunation.

How the Earth is presented towards the Sun, during
the Eclipse, and the path of the Moon's Shadow.

This tabular summary shows:

1st. That the quarters arrange themselves in an entirely dif ferent order under the different heads, with the single exception of the second being the quarter of greatest average rainfall and also of greatest average wet-day rate of rain.

2nd. That the least average rainfall was in the quarter immediately preceding the new moon, instead of being, as Sir J. Herschel supposes, about the full moon.

3rd. That the maximum number of wet days was in the third quarter, and the minimum in the first; thus differing in every particular from the results stated by M. Arago to have been obtained in Germany and Paris, on the one hand, and in the south of France on the other, which, as we have seen, differed from one another.

This discussion may be appropriately closed, perhaps, by echoing Arago's remark, that "the question requires to be examined afresh."

V. THE APPROACHING TOTAL SOLAR ECLIPSE.

By R. A PROCTOR, F.R.A.S., &c.

The

THE eclipse of next December is less remarkable in many important respects than the two total solar eclipses now commonly known as the Indian and American eclipses of 1868 and 1869. The former of these was distinguished among all the eclipses of recent times by the exceptional extent to which the lunar disc overlapped, during central totality, the concealed disc of the sun. For more than six minutes at some stations no direct solar light was visible. eclipse of last year was not distinguished in this particular way, though the duration of totality-at some stations exceeding four minutes-was far from inconsiderable. What rendered the American eclipse so extremely important, even more important than the Indian one, was the fact that a large proportion of the track of the moon's shadow lay across a region dotted over with well-armed observatories. It is probable that on no previous occasion has so large an array of practised observers been employed in scrutinizing the phenomena of a total eclipse; and it is absolutely certain that so many appliances had never before been employed to render the researches of the observers effective.

In both respects the approaching eclipse is less important. The greatest duration of total obscuration will be but 2 m. 11 s.; and the track of the moon's shadow only skirts the region within which the principal European observatories are situated. In fact, the only parts of Europe traversed by the shadow are the southern provinces of Spain and Portugal, Sicily, the southern extremity of Italy, and