nomer

on the merits of Mr. Airy's paper, as affording sufficient evidence of the justice and importance of the Astronomer-Royal's views. The chief circumstance to be noted, however, is that the new inquiry, though based on novel considerations and conducted in a novel manner, yet led to precisely the same result as former researches. Not only was the original estimate of the direction of the sun's motion confirmed by the Astroer-Royal, but the velocity with which the sun is journeying through space came out from his figures and formulæ appreciably unchanged. Mr. Dunkin, of the Greenwich Observatory, to whose skill the Astronomer-Royal had entrusted the laborious and difficult processes involved in the application of the new method to 1,167 stars, remarks on this head that the mean of the values obtained by Otto Struve and by himself differs so little from either, that we may look upon it as fairly representing the annual motion of the sun, so far as any result can be obtained by the use of the apparent proper motions of the stars.' The sun's velocity, according to both methods, is found to be about 150 millions of miles per annum.

It would seem, then, that the problem of the sun's motion in space has been placed in the category of settled questions. With so many solutions, various in method, founded on different sets of stars, and carried out by the most skilful mathematicians to all but identical results, no one can reasonably doubt, it might be thought, that the sun's motion is such as has been stated by our astronomers.

And yet, singularly enough, this last and most satisfactory of all the solutions which the problem has received has introduced an element of doubt into the question which it is impossible to overlook. The Astronomer-Royal was not content

with the mere solution of the problem, but persisted, after the obstinately inquisitive manner usual amongst astronomers, in applying all manner of tests to the result. Among these tests was one of a most interesting character.

It was argued that if our sun's motion is reflected so clearly amongst the drifting stars, we ought to find the amount of stardrift largely diminished when the full correction is made for the sun's motion. Accordingly the experiment was tried. Every star of the 1,167 in Mr. Main's list was carefully set drifting in a direction precisely opposite to that due to the sun's motion; the effect being intended to be such as to correspond to that which would really take place if the sun were brought to rest. I say intended, because the doubts which hang over the question of the stars' distances come in here again to perplex the question. However, the best was done which the circumstances admitted of. The estimates of the elder Struve were adopted, and the immensely laborious work of correction was carried out to its completion by Mr. Dunkin.

The result was not such as was to have been expected. I premise that, in this case, the sum of the squares of the uncorrected motions has to be brought into comparison with the sum of the squares of the corrected motions. The former sum, then, was found to be 1420251, the latter 136 4915. The correction, instead of being important as was anticipated, is less than a twenty-fifth part of the whole.

Sir John Herschel, commenting on this result, remarks: 'No one need be surprised at this. If the sun move in space, why not also the stars? and if so, it would be manifestly absurd to expect that any movement could be assigned to the sun by any system of calcula tion which should account for more

than a very small portion of the totality of the observed displacements. But what is indeed astonishing in the whole affair is that among all this chaotic heap of miscellaneous movement, among all this drift of cosmical atoms, of the laws of whose motions we know absolutely nothing, it should be possible to place the finger on one small portion of the sum total, to all appearance undistinguishably mixed up with the rest, and to declare with full assurance that this particular portion of the whole is due to the proper motion of our own system.'

With all deference to an authority so distinguished, I must venture to express my doubts of the correctness of the opinion I have just quoted. Sir John Herschel has omitted to consider that the number of bodies affected by their own proper motion has nothing whatever to do with the effects of the correction under discussion; because every one of those bodies is affected by the sun's proper motion. The largeness of the number is as effective one way as the other. This general consideration suffices to throw Herschel's conclusion into doubt; but in reality the question is one for mathematical discussion, and when this has been applied it becomes certain that his conclusion is

erroneous.

Now I could not present here the processes by which I have established the view that a larger correction was to have been ex

pected. But I can indicate a line of reasoning which will exhibit very clearly the probability of that view. Suppose there are two stars in that part of the heavens where the sun's motion produces the greatest change of position, and that one star is moving in the same direction as the sun and with the same velocity, the other in the contrary direction. It is obvious that the first has its apparent motion re

duced to rest by the effect of the sun's motion; the second has its motion doubled. Thus, instead of the motions 1, 1 (as we may count them), which would be observed if the sun were at rest, we get the motions o, 2; and instead of the squares I, 1, we get the squares 0, 4, an increase in all of 2. Now, any one familiar with the elements of geometry will find that, even if the two stars had opposite motions along some other line than the one parallel to the sun's path, so that (situated as the stars are supposed to be) the squares of their own motions would be less than 1, 1, the increase due to the sun's motion would still be 2. Hence, if we suppose a large number of stars to be moving in all directions from a point situated as supposed, every pair of opposite motions would have the sum of its squares increased by 2; whereas the original sum would never exceed 2, and would commonly be less than 2. Hence the effect of the sun's motion would be to more than double the sum of the squares for stars in this particular part of the heavens. For stars situated elsewhere the sun's motion would be less effective. Without entering further into explanations of the probable effect of the sun's motion for stars all over the sky, I may mention, as the result of an exact mathematical discussion of the subject, that, on the whole, the sun's motion ought exactly to double the sum of the squares. words, the correction due to the sun's motion ought to reduce the sum of squares by one-half. We have seen that the actual reduction is less than one twenty-fifth.

In other

We have, then, only three explanations to choose from. Either the sun's motion is considerably less than the average motions of the stars; or stars are moving according to some law which tends to mask the effect of the motion of the sun, which is but one among

their number; or, lastly, the assumptions which have been made respecting the stars' distances (as judged by their brightness) are wholly incorrect.

We have seen, at the very beginning of this inquiry, that the sun's motion cannot be looked upon as small in comparison with the average motions of the stars. Thus the first explanation must be dismissed. As to the second, it is clearly improbable in the highest degree that that should be the real explanation of our difficulty; for besides that a law, to be so effective in masking the effect of the sun's motion, ought to exhibit itself by obvious relations among the stellar movements, we have the fact that the sun's motion is not masked when estimated in other ways. It is only in one sense that it is masked at all, and its effect, viewed in that light, depends entirely on the assumptions which have been made respecting the stars' distances. We are led, then, to the third explanation, which involves the very conclusion which I have endeavoured to establish above on other grounds-the conclusion, namely, that the distances of the stars are not to be estimated by the stars' apparent brightness.

It only remains to mention that the same amount of change in our estimates of the stars' distances, which the former mode of inquiry suggested, appears to correspond closely with the amount required

to make the corrections due to the sun's motions sufficiently effective. I can only speak positively, however, for the stars of the lower magnitudes, because the correction due to the sun's motion is in reality not a correction at all for stars of the second and third magnitudes, since it actually increases the sum of the squares. My formulæ do not enable me to deal with a negative correction, though they show very conclusively that such a correction can only result from mistaken assumptions. The fact that such corrections exist is interesting also as confirming a result already indicated in this paper-viz. that the stars of the first three magnitudes must be taken together in inquiries of this sort.

The result of my examination of the subject seems to prove that we can no longer assume the stars to be all of the same or nearly the same magnitude, but that, on the contrary, their differences of magnitude are so important as to constitute the chief explanation of their differences of brilliancy. But this conclusion, which seems to me indisputable, in no sense contravenes the conclusions which astronomers have formed respecting the sun's motion through space, though it throws considerable doubt over the accuracy of the estimates which have been formed of his velocity, in miles per annum.

of Europe without acquiring any intimate knowledge-philological, ethnical, or otherwise of the inhabitants of any other nationality of the same quarter of the globe. Strange as it may appear, it is equally possible for an Anglo-Indian to live for years in India and remain almost totally ignorant of the language, history, manners and customs of the natives among whom he dwells, and whom it is his lot to see every day of his existence.

To the great mass of Englishmen British India, from Peshawur to Cape Comorin, is ethnically and physically one, just as is Spain or Holland and its two hundred millions of population are Hindoos and Mahommedans of one or, at the most, two races. There may be added to this much information a vague impression that the Hindoos are divided into classes by ordinances of caste; and it may even be known that, at some time or other, the Aryans (whoever they may have been) and the Mahommedans entered India from foreign parts, and made themselves conspicuous in the History of Hindostan. Many AngloIndians, so taught, fail during a lifetime in India to add very much to their knowledge of the subject. Men of business may learn something of the worst sides of the native character; they may discover that the Asiatic is not a model of probity, whether he be a denizen of Bengal, Madras, Bombay, or any of the northern provinces-that the argumentum ad verecundiam does not apply to dealings with the native of India as it does where Europeans are concerned and that the conquered

people are not nicely scrupulous in

their conquerors and each other. But they may fail to penetrate the character of the people at any other point: and what little they discover of the most superficial characteristics and customs of those around them they erroneously assume to be equally applicable to all the people of India-predicating for the whole what in truth only applies to a part.

It is a logical deduction that knowledge of a people's character and mode of life must be preceded by more or less intimacy with the language of that people: and when it is stated that hundreds (thousands would possibly be within the mark) of Anglo-Indians never learn anything of the native dialects, it will be readily intelligible that their ignorance should be considerable upon all those points as to which information can only come of intercourse. To many Anglo-Indians there never occurs the necessity of speaking any native tongue. The merchants, bankers, and other nonofficials settled down in a provincial capital can transact their business without employing any other than Anglo-Saxon speech. English is the non-official language; and every native employé above the rank of messenger speaks it as a matter of course. In Madras and Bombay those servants who come into contact with the master speak English (at least it is more like English than any other language) and act as interpreters for the remainder of the household and the hangers-on of the domestic world. In other parts of India English-speaking servants can be procured with little difficulty. And, at the worst, the non-official of the provincial capital only requires such

a limited vocabulary as will enable him to order dinner.

So, too, the exigencies that drive the officer of a British regiment to depart from his mother-tongue are few and far between. He generally has a servant who speaks English: the mess-waiters are men of the regiment; and, if he know enough of the vernacular to ask his way back to cantonments or to his camp when he has gone into the interior and missed his road, he knows enough to meet any emergency likely to befall him. There are, of course, many officers of Her Majesty's British forces who learn more than the minimum positively required of them; but there are not a few who learn less, and are hopelessly stranded when they make the slightest colloquial attempt. Nor is this unnatural. The British officer is in India only for a time (unless he exchange he may not be there longer than five to ten years); that time may be broken by frequent absences on leave, and he cannot anticipate a return to that country after his regiment leaves it. With a sufficient show of reason he argues that it is useless importing the Oordoo or Hindoo language into England, where it can only be employed to the bewilderment of a Lascar crossing-sweeper whose native dialect is Bengali or Tamil, and from whose linguistic répertoire Oordoo and Hindoo have been wholly omitted; and, considering the matter from this practical point of view, he leaves the languages to the regimental interpreter.

To some of the expedients adopted and errors made by Anglo-Indians who have failed to master the vernacular, has been accorded a worldwide fame (Anglo-Indian) as great as was ever won by the bons mots of Sheridan or the best utterances of Rochefoucauld. Almost as historical a personage as Warren Hastings is that Calcutta merchant who, wishing to ask what pie he had before

him on the dinner table, demanded of his Khansamah (in Hindoostanee) whose pie it was, and was duly informed that it was his own. Great, too, is the young officer who, being asked in the course of a colloquial examination how he would express himself in the vernacular if he wished his syce to take his buggy under the shade of a tree, escaped out of the difficulty by saying, that he should go under the tree himself and beckon to the syce to bring the buggy there. And, were it necessary, the list of those who have established for themselves celebrity after the same fashion might be further added to.

To those who take the trouble to inquire into the matter, it is known that the millions of British India are divided into peoples differing as widely as-sometimes more widely than the various peoples of Europe. The Aryans, wherever they are found, are of one family, though Tartar and Afghan conquerors have often corrupted the stream of Aryan descent. The Mahommedans, with the exception of the few whose progenitors were forcibly converted to Islamism when the crescent first descended upon the plains of India, are the descendants of those warriors who built up the throne of Delhi and made of Hindostan an empire for their mogul. But among the aborigines there are many tribes which are ethnically and philologically divided. There are tribes closely related to the Malay and Chinese races, and there others that are wholly unconnected with those branches of the human race. The number of aboriginal dialects spoken throughout India has been computed to exceed two hundred, and none of these has any marked affinity to any other. the insignificant Naga country near Assam thirty distinct dialects have been detected, and in estimating the number for all India at about two hundred, the error, if there be any,

are

In