as the chief of the Observatory and some of his assistants could entertain any doubt on that point. Probably it was intended by the words just quoted to imply simply that some of the observations were made for the purpose of illustrating the principle of the method. We are not to suppose that on a point so simple the Greenwich observers have been in any sort of doubt.

At first their results were not very satisfactory. The difficulties which had for a long time foiled Huggins, and which Secchi has never been able to master, rendered the first Greenwich measures of stellar motions in the line of sight wildly inconsistent, not only with Huggins's results, but with each other. Secchi was not slow to note this, and a short time ago he renewed his objections to the new method of observation, pointing and illustrating his objections by referring to the discrepancies among the Greenwich results. But recently a fresh series of results has been published, showing that the observers at Greenwich have succeeded in mastering some at least among the difficulties which they had before experienced. The measurements of starmotions showed now a satisfactory agreement with Huggins's results, and their range of divergence among themselves was greatly reduced. The chief interest of the new results, however, lay in the observations made upon bodies known to be in motion in the line of sight at rates already measured. These observations, though not wanted as tests of the accuracy of the principle, were very necessary as tests of the accuracy of the instruments used in applying it. It is here and thus that Secchi's objections alone required to be met, and here and thus they have been thoroughly disposed of. Let us consider what means exist within the solar system for thus testing the new method.

The earth travels along in her orbit at the rate of about 18 miles in every second of time. Not to enter into niceties which could only properly be dealt with mathematically, it may be said that with this full velocity she is at times approaching the remoter planets of the system, and at times receding from them; so that here at once is a range of difference amounting to about 37 miles per second, and fairly within the power of the new

method of observation. For it matters nothing, so far as the new method is concerned, whether the earth is approaching another orb by her motion, or that orb approaching by its own motion. Again, the planet Venus travels at the rate of about 211⁄2 miles per second, but as the earth travels only 3 miles a second less swiftly, and the same way round, only a small portion of Venus's motion ever appears as a motion of approach towards or recession from the earth. Still Venus is sometimes approaching and sometimes receding from the earth, at a rate of more than 8 miles per second. Her light is much brighter than that of Jupiter or Saturn, and accordingly this smaller rate of motion would be probably more easily recognized than the greater rate at which the giant planets are sometimes approaching and at other times receding from the earth. the Greenwich observers seem to have confined their attention to Venus, so far as motions of planets in the line of sight are concerned. The moon, as a body which keeps always at nearly the same distance from us, would of course be the last in the world to be selected to give positive evidence in favor of the new method; but she serves to afford a useful test of the accuracy of the instruments employed. If when these were applied to her they gave evidence of motions of recession or approach at the rate of several miles per second, when we know as a matter of fact that the moon's distance never * varies by more than 30,000 miles during the lunar month, and her rate of approach or recession thus averaging about one-fiftieth part of a mile per second, discredit would be thrown on the new methodnot, indeed, as regards its principle, which no competent reasoner can for a moment question, but as regards the possibility of practically applying it with our present instrumental means.

Observations have been made at Greenwich, both on Venus and on the moon, by the new method, with results entirely satisfactory. The method shows that Venus is receding when she is known to be receding, and that she is

* It varies more in some months than in others, as the moon's orbit changes in shape under the various perturbing influences to which she is subject.

It

approaching when she is known to be approaching; and the method shows no signs of approach or recession in the moon's case, and is thus in satisfactory agreement with the known facts. Of course these results are open to the objection that the observers have known beforehand what to expect, and that expectation often deceives the mind, especially in cases where the thing to be observed is not at all easy to recognize. It will presently be seen that the new method has been more satisfactorily tested, in this respect, in other ways. may be partly due to the effect of expectation that in the case of Venus the motions of approach and recession, tested by the new method, have always been somewhat too great. A part of the excess may be due to the use of the measure of the sun's distance, and therefore the measures of the dimensions of the solar system, in vogue before the recent transit. These measures fall short to some degree of those which result from the observations made in December, 1874, on Venus in transit, the sun's distance being estimated at about 91,400,ooo miles instead of 92,000,000 miles, which would seem to be nearer the real distance. Of course all the motions within the solar system would be correspondingly under-estimated. On the other hand, the new method would give all velocities with absolute correctness if instrumental difficulties could be overcome. The difference between the real velocities of Venus approaching and receding, and those calculated according to the present inexact estimate of the sun's distance, is however much less than the observed discrepancy, doubtless due to the difficulties involved in the application of this most difficult method. I note the point, chiefly for the sake of mentioning the circumstance that theoretically the method affords a new means of measuring the dimensions of the solar system. Whensoever the practical application of the method has been so far improved that the rate of approach or recession of Venus, or Mercury, or Jupiter, or Saturn (any one of these planets) can be determined on any occasion, with great nicety, we can at once infer the sun's distance with corresponding exactness. Considering that the method has not been invented ten years (setting

aside Doppler's first vague ideas respecting it), and that spectroscopic analysis. as a method of exact observation is as yet little more than a quarter of a century old, we may fairly hope that in the years to come the new method, already successfully applied to measure motions of recession and approach at the rate of 20 or 30 miles per second, will be employed successfully in measuring much smaller velocities. Then will it give us a new method of measuring the great base-line of astronomical surveying-the distance of our world from the centre of the solar system.

That this will one day happen is rendered highly probable, in my opinion, by the successes next to be related.

Besides the motions of the planets around the sun, there are their motions of rotation, and the rotation of the sun himself upon his axis. Some among these turning motions are sufficiently rapid to be dealt with by the new method. The most rapid rotational motion with which we are acquainted from actual observation is that of the planet Jupiter. The circuit of his equator amounts to about 267,000 miles, and he turns once on his axis in a few minutes less than ten hours, so that his equatorial surface travels at the rate of about 26,700 miles an hour, or nearly 7 miles per second. Thus between the advancing and retreating sides of the equator there is a difference of motion in the line of sight amounting to nearly 15 miles. But this is not all. Jupiter shines by reflecting sunlight. Now it is easily seen that where his turning equator meets the waves of light from the sun, these are shortened, in the same sense that waves are shortened for a swimmer travelling to meet them, while these waves, already shortened in this way, are further shortened when starting from the same advancing surface of Jupiter, on their journey to us after reflection. In this way the shortening of the waves is doubled, at least when the earth is so placed that Jupiter lies in the same direction from us as from the sun, the very time, in fact, when Jupiter is most favorably placed for ordinary observation or at his highest due south, when the sun is at his lowest below the northern horizon

that is, at midnight. The lengthening of the waves is similarly doubled at this

most favorable time for observation; and the actual difference between the motion of the two sides of Jupiter's equator being nearly 15 miles per second, the effect on the light-waves is equivalent to that due to a difference of nearly 30 miles per second. Thus the new method may fairly be expected to indicate Jupiter's motion of rotation. The Greenwich observers have succeeded in applying it, though Jupiter has not been favorably situated for observation. Only on one occasion, says Sir G. Airy, was the spectrum of Jupiter "seen fairly well," and on that occasion " measures were obtained which gave a result in remarkable agreement with the calculated value." It may well be hoped that when in the course of a few years Jupiter returns to that part of his course where he rises high above the horizon, shining more brightly and through a less perturbed air, the new method will be still more successfully applied. We may even hope to see it extended to Saturn, not merely to confirm the measures already made of Saturn's rotation, but to resolve the doubts which exist as to the rotation of Saturn's ring-system.

Lastly, there remains the rotation of the sun, a movement much more difficult to detect by the new method, because the actual rate of motion even at the sun's equator amounts only to about 1 mile per second.

In dealing with this very difficult task, the hardest which spectroscopists have yet attempted, the Greenwich observers have achieved an undoubted success; but unfortunately for them, though fortunately for science, another observatory, far smaller and of much less celebrity, has at the critical moment achieved success still more complete.

The astronomers at our national observatory have been able to recognize by the new method the turning motion of the sun upon his axis. And here we have not, as in the case of Venus, to record merely that the observers have seen what they expected to see because of the known motion of the sun. "Particular care was taken," says Airy, "to avoid any bias from previous knowledge of the direction in which a displacement" (of the spectral lines) " was to be expected," was to be expected," the side of the sun under observation

not being known by the observer until after the observation was completed.

Let

But Professor Young, at Dartmouth College, Hanover, N. H., has done much more than merely obtain evidence by the new method that the sun is rotating as we already knew. He has succeeded so perfectly in mastering the instrumental and observational difficulties, as absolutely to be able to rely on his measurement (as distinguished from the mere recognition) of the sun's motion of rotation. The manner in which he has extended the powers of ordinary spectroscopic analysis, cannot very readily be described in these pages, simply because the principles on which the extension depends require for their complete description a reference to mathematical considerations of some complexity. it be simply noted that what is called the diffractive spectrum, obtained by using a finely-lined plate, results from the dispersive action of such a plate, or grating as it is technically called, and this dispersive power can be readily combined with that of a spectroscope of the ordinary kind. Now Dr. Rutherfurd of New York has succeeded in ruling so many thousand lines on glass within the breadth of a single inch as to produce a grating of high dispersive power. Availing himself of this beautiful extension of spectroscopic powers, Professor Young has succeeded in recognizing effects of much smaller motions of recession and approach than had before been observable by the new method. He has thus been able to measure the rotation-rate of the sun's equatorial regions. His result exceeds considerably that inferred from the telescopic observation of the solar spots. For whereas from the motion of the spots a rotation-rate of about 14 miles per second has been calculated for the sun's equator, Professor Young obtains from his spectroscopic observations a rate of rather more than 1 miles, or about 300 yards per second more than the telescopic rate.

If Young had been measuring the motion of the same matter which is observed with the telescope, there could of course be no doubt that the telescope was right and the spectroscope wrong. We might add a few yards per second for the probably greater distance of the

sun resulting from recent transit observations. For of course with an increase in our estimate of the sun's distance there comes an increase in our estimate of the sun's dimensions, and of the velocity of the rotational motion of his surface; but only about 12 yards per second could be allowed on this account, the rest would have to be regarded as an error due to the difficulties involved in the spectroscopic method. But in reality the telescopist and the spectroscopist observe different things in determining by their respective methods the sun's motion of rotation. The former observes the motion of the spots, belonging to the sun's visible surface; the latter observes the motion of the glowing vapors outside that surface, for it is from these vapors, not from the surface of the sun, that the dark lines of the spectrum proceed. Now so confident is Professor Young of the accuracy of his spectroscopic observations, that he is prepared to regard the seeming difference of velocity between the atmosphere and surface of the sun as real. He believes that "the solar atmosphere really sweeps forward over the underlying surface, in the same way that the equatorial regions outstrip the other parts of the sun's surface." This inferThis infer ence, important and interesting in itself, is far more important in what it involves. For if we can accept it, it follows that the spectroscopic method of measuring the velocity of motions in the line of sight is competent, under favorable conditions, to obtain results accurate within a few hundred yards per second, or 10 or 12 miles per minute. If this shall really prove to be true for the method now, less than nine years after it was first successfully applied, what may we not hope from the method in future years? Spectroscopic analysis itself is

in its infancy, and this method is but a recent application of spectroscopy. A century or so hence astronomers will smile (though not disdainfully) at these feeble efforts, much as we smile now in contemplating the puny telescopes with which Galileo and his contemporaries studied the star-depths. And we may well believe that largely as the knowledge gained by telescopists in our own time surpasses that which Galileo obtained, so will spectroscopists a few generations hence have gained a far wider and deeper insight into the constitution and movements of the stellar universe than the spectroscopists of our own day dare even hope to attain. I venture. confidently to predict that, with that insight, astronomers will recognize in the universe of stars a variety of structure, a complexity of arrangement, an abundance of every form of cosmical vitality, such as I have been led by other considerations to suggest, not the mere cloven lamina of uniformly scattered stars more or less resembling our sun, and all in nearly the same stage of cosmical development, which the books of astronomy not many years since agreed in describing. The history of astronomical progress does not render it probable that the reasoning already advanced, though in reality demonstrative, will convince the generality of science-students until direct and easily understood observations have shown the real nature of the constitution of that part of the universe over which astronomical survey extends. But the evidence already obtained, though its thorough analysis may be "caviare to the general," suffices to show the real nature of the relations which one day will come within the direct scope of astronomical observation.-Contemporary Review.

I.

IN admitting into the pages of the Nineteenth Century a narrative of an

With the exception of the introductory remarks, the following paper is wholly composed of extracts from the author's note-book, written afloat and for the most part at sea.

amateur voyage of circumnavigation, I fear that the Editor runs a risk of descending into a sphere too narrow in its scope to deserve the attention of a large public.

But as he decides to run that risk I make no further apology, and address

myself at once to the task which I have been requested to undertake. I commence with a general outline of the voyage, and shall subsequently fill in the details of the picture, which, unless connected together at the commencement by a slight sketch of the whole cruise, would be seen in a disjointed and fragmentary aspect.

The expedition was in some respects unprecedented; and the most exceptional feature was the little company of passengers. They included Mrs. Brassey and our four children. The youngest was less than two years of age, and has returned to England in robust health. A voyage of circumnavigation is an ordinary undertaking for a professional seaman; but it was no inconsiderable effort for a lady to exchange the luxuries of an English home for an uneasy residence of eleven months on the rolling sea. And what shall be said of the nurses? True daughters of their Scandinavian forefathers, they accepted the unusual and trying conditions of their sea life with undaunted spirit, and showed no symptoms either of fear or discontent from the day of their departure to the hour of their final disembarkation. A circumnavigation of 35,400 miles has never before been made in the short period of 46 weeks, from which must be deducted 112 days of well-earned repose in harbor. We had, it is true, the advantage of steam, without which such a performance would have been an impossibility; but we travelled 20,517 miles under sail alone, and the consumption of coal has not exceeded 400 tons.

The Sunbeam' sailed from Cowes on the 6th of July, 1876, put into Torbay on the following day, resumed her voyage on the 8th, and reached Madeira on the 16th of July. Strong winds were experienced in the Channel, and a fresh gale from the north-east off Cape Finisterre. South of the latitude of Lisbon calms prevailed. In this stage of the voyage 353 miles were traversed under steam, and 886 miles under sail.

Leaving Madeira on the 20th of July, we called at Orotava, for the ascent of the Peak of Teneriffe, and at Tarafal Bay, in the island of San Antonio, one of the Cape de Verdes, for provisions, arriving at Rio de Janeiro on the 17th of August. We sailed before the north

east trades from Teneriffe to Tarafal Bay, and thence pursued our voyage across the Atlantic to Rio.

The Sunbeam' again put to sea on the 5th of September, and in six days reached Montevideo. On the 8th and 9th a gale blew from the north-east; the distances sailed under reefed canvas on these two days being 243 and 270 knots respectively. During our stay in the River Plate we spent a fortnight at Buenos Ayres, and made excursions to Rosario and Cordova, and to Azul, on the southern frontier; we afterwards visited Ensenada.

The voyage was resumed on the 28th of September, and on the 6th of October we arrived at Sandy Point, in the Straits of Magellan. On this passage we rescued a crew of fifteen hands from the barque 'Monks' Haven,' bound from Cardiff to Valparaiso with a cargo of smelting coals. On the 2nd of October we encountered a gale from the southwest, but escaped its full effects by closing with the coast of Patagonia.

The voyage was continued through the Straits of Magellan and Smyth's Channel. It was our happy fortune to see the magnificent mountains of those 'stern and wild' regions in most auspicious weather. The distance from the eastern entrance to the Straits of Magellan to the northern outlet from Smyth's Channel into the Gulf of Penas was 659 miles. We made the passage under steam in seventy-six hours. Aided by the admirable charts from the surveys of Captain King, Admiral Fitzroy, and Captain Mayne, C.B., we were enabled to navigate these intricate channels at full speed, and find well-sheltered anchorages every night.

Lota was our first port on the coast of Chili, and on the 21st of October we reached Valparaiso. After a stay of nine days in that busy but ill-protected harbor, we proceeded on our long and lonely voyage of 12,333 miles across the Pacific to Yokohama. We touched at Bow Island in the Low Archipelago, at Maitea and Tahiti in the Society Islands, at Hawaii and Honolulu, in the Sandwich Islands, sighted Assumption, an isolated extinct volcano in the Ladrones, on the 21st of January, and arrived at Yokohama on the 29th. We had made the passage from Valparaiso in seventy-two days