The Indian Meteorological Department is making preparations to carry out a portion of this work; and will undoubtedly do the best it can single-handed with its limited means. It cannot do the work fully and as it ought to be done. It can do nothing which requires authoritative control over the remaining meteorological systems in the Indo-oceanic field. It is collecting information from those who are willing to supply it, and will utilise it for its special purposes.

It is evident the work can only be carried out fully by the co-operation of the various systems subject to limited control by a central office with acknowledged imperial or general authority behind it. The most important part of the work from the standpoint of the science of meteorology is the comparison and discussion of the whole body of observations. The constitution, position, and authority of the central office is hence of the greatest importance. It is quite certain that none of the meteorological systems directly concerned can provide such a central office. If the work is to be carried out fully and systematically it can only be arranged for in England, and by the English Government assuming the general direction and control. At the present time a section of the English Meteorological Office is devoted to the study of oceanic meteorology for the information of mariners. Another section should be created for the study of imperial meteorology for the benefit of its dependencies and colonies. I have reason to believe that the Government of India would contribute its share towards the cost of this extension of work.

In the preceding remarks are given the chief reasons for an important extension of work now in progress in the Indian Meteorological Department, an extension which can only be carried out imperfectly by that Department, but which could be performed with most valuable scientific results by the co-ordination of the labours of the weather bureaus concerned, with a central institution or investigating office in England under Government control.

Perhaps I may be permitted, from my Indian experience, to add some general remarks bearing on the methods and progress of meteorological inquiry.

In India the collection and publication of accurate current data relating to rainfall and temperature is required for the information of Government in its various Departments. The collection and examination of pressure and wind data by a central office with a view to the issue of storm and flood warnings is equally necessary. This work may, perhaps, be described as pertaining to descriptive or economic meteorology.

Economic meteorology, so long as it deals only with actual facts of observation, is not a science. Forecasts belong to the same department or branch of meteorology. They may be based on scientific theory and be obtained by scientific methods or the utilisation of empirical knowledge. The latter method is probably sufficient for by far the greater part of short-period forecast work, but the final development of that work and the preparation of long-period forecasts require the application of exact scientific methods and knowledge. And it is, perhaps, not too much to say that the extension of the range or period of forecasts is a measure of the progress of meteorology as a science. India, by the simplicity and massiveness of its meteorological changes (and perhaps Australia and Africa), appears to be best suited for the earliest experiments in this work.

India is, however, poor, not only in material wealth and capital as compared with England, but also in the appliances and means of scientific investigation, and hence looks to England for assistance and guidance in scientific matters. Unfortunately, England lags behind, not only the United States and Germany, but even behind India, in the important field of scientific meteorological inquiry. It will suffice to give a single illustration of the anomalous and inferior position which England takes in such matters.

All meteorologists and scientific men generally are agreed that the exploration of the middle and upper atmosphere by any available means-e.g., kites, balloons, &c.-is of the utmost importance at the present stage of meteorological inquiry. The United States, France, and Germany have taken up the work vigorously. The English Meteorological Office is unable, for want of funds, to share or take any part in the work. The force of scientific and public opinion is apparently

TRANSACTIONS OF SECTION A.

powerless to move the English Government to grant an extra five hundred pounds The annually for this work. The English Government, on the other hand, some time ago suggested that the Indian Meteorological Department should assist. Government of India, recognising the importance of the work, has provided the funds and sanctioned the arrangements necessary in order that its Meteorological Department may march with the most progressive nations in this investigation.

resources.

India has no body of voluntary observers or independent scientific workers and investigators. Whatever is required to be done to extend practical and theoretical meteorology can only be effected by the Government Department to which that work is assigned, with the sanction and at the cost of the Government -which naturally considers chiefly its practical wants in relation to its limited It is, from one point of view, a painful if not quite an unexpected experience to me, on my retirement, to find that the Government of India is, in its attitude towards meteorological inquiry, more advanced, more liberal and far-sighted than the English Government, and that England has not yet taken up seriously the work of scientific meteorological investigation. There are undoubtedly too many observations and too little serious discussion of observations. The time has arrived when investigation should go hand in hand with accurate observation, and should direct and suggest the work of observation, and also that the sciences directly related to meteorology should be considered concurrently with it. There are undoubtedly definite relations between certain classes of solar phenomena and phenomena of terrestrial magnetism. The probability of definite relations between solar and terrestrial meteorological phenomena is also generally admitted.

Data for the determination of these relations are being rapidly accumulated, and numerous problems connected therewith are waiting and ripe for investigation. They are too large and complex to be undertaken by present English methods, and can only be attacked by a body of trained investigators under arrangements securing the continuity of method and thought requisite for the prolonged systematic inquiry gradually leading up to their complete solution.

It would hence be desirable to enlarge the scope of the central institution I have suggested, so as to include in its field of labour the investigation of the relation between solar and terrestrial meteorology and magnetism, so far as they can be solved by the comparison of the observations of the British Empire.

The central institution would thus have large and definite fields of work and most interesting problems for investigation. It would hence contribute towards the formation of a body of scientific meteorological investigators adequate to the importance and wants of the empire, and be of the highest educational as well as scientific value.

My predecessor in this position, Dr. Shaw, the head of the English Meteorological Office, made some remarks in his Address last year which deserve repetition in connection with this idea. He said: 'The British Empire stands to gain more by scientific knowledge, and to lose more by unscientific knowledge, of the matter than any other country. It should from its position be the most important agency for promoting the advance of meteorological science-in the first place because it possesses such admirable varying fields of observation, and in the second place because with due encouragement British intellect may achieve as fruitful results in this as in other fields of investigation.'

The establishment of the central institution as suggested above would provide a remedy for the defects pointed out by Dr. Shaw. The reorganisation of the English Meteorological Office is, I believe, under consideration. Is it too much to hope that a strong expression of opinion on the part of the British Association, and the influence of the learned University at which its present meeting is held, would induce the English Government to spend an additional 5,000l. or 10,0007. annually for the promotion of meteorological investigation and the establishment of a central Imperial institution in London in connection with its Meteorological Office?

The following Papers were read :—

1. The Spectra of Sun-spots.

By the Rev. A. L. CORTIE, S.J., F.R.A.S.

The paper contains a reduction of all the observations of sun-spot spectra taken at Stonyhurst during the years 1883-1901 with a 12-prism spectroscope attached to either the 8-inch or 15-inch equatorial. A discussion of the observations of the spectra of ninety sun-spots, observed during the period 1883-1889, appeared in the Memoirs R.A.S.' vol. 1, and of twenty-four other spots observed in the period 1890-1901 in the Monthly Notices R.A.S.' vol. İxiii. No. 8. All the observations have been taken by the same observer, and are not restricted to a few lines, but on each occasion some particular region of the spectrum between B and D has been selected for detailed study, after a general view of this part of the spectrum had been secured for determining the most widened lines. The earlier observations of the widened lines were catalogued according to Angström's wave-length numbers, as corrected in the British Association 'Catalogue of the Oscillation-frequencies of the Solar Rays' (1878); the later observations according to Rowland's numbers. The present catalogue of 346 widened lines between wave-lengths 5884-03 and 6867 46, which combines all former lists, is based on Rowland's numbers, and contains 5486 individual observations.

The chief phenomena in the spectra of sun-spots are, as regards the general absorption, a want of uniformity of blackness in various regions of the spectrum sometimes observed, and, as regards the selective line absorption, the widening of lines, darkening of lines without widening, displacement of lines, obliteration of lines, extension of the widening through the penumbra, reversal of lines, hazy fringes to lines, and spot-bands. The following tables contain, the one, the mean relative widening of the lines of the chief elements identified, and the other, a list of the most widened lines. The numbers for relative widening are estimated in terms of the normal width of the line multiplied by the factor 10, and for intensities are taken from Rowland's Catalogue, where 1 is a line just clearly visible, and successive zeros indicate increasing degrees of faintness.

The tables show the importance of faint lines of vanadium and titanium in the sun-spot spectra (Monthly Notices R.A.S.' vol. lviii. No. 7). These faint lines have been always, and at all times of the sun-spot period, among the most widened lines (loc. cit. vol. xlix. No. 8; vol. lxii. No. 7). The observations give no evidence of a crossing of the most widened lines at an epoch between sunspot maximum and minimum. They show, however, that the iron lines are more affected in minimum than in maximum spots; no conclusion can be drawn as to a difference of character and temperature between maximum and minimum spots from the behaviour of such faint widened lines. The iron lines brightened in the chromosphere, which are mostly arc lines, are not differently affected in sun-spots from lines not brightened. The widening of some oxygen lines in sun-spots in the a band seems to be a real phenomenon, the single hydrogen line (C) is generally thinned, and almost reversed over spots, and frequently reversed and distorted in their immediate neighbourhood. If oxygen and hydrogen are present in sun-spots, water-vapour might be formed over them. Spot-bands sometimes seen (loc. cit. vol. xlvii. No. 1) are a probable witness to a reduction of temperature sufficient for the formation of compounds. But the widened lines accredited to water-vapour occur generally in crowded parts of the spectrum, so that the widening may be really due to faint solar lines in juxtaposition with them. The predominance of vanadium and titanium in sun-spots is important in view of Mr. Fowler's recent identification of the flutings in Secchi's third-type stars as due to titanium or a titanium compound, and Sir Norman Lockyer's matching of the lines intensified in the spectrum of Arcturus with lines of the same element. Professor Hale has also shown that many of the lines in the fourth-type stars are coincident with lines observed as widened in sun-spots by Mr. Maunder and myself.

TABLE I.-Relative Widening of the Lines of each Element.

TABLE II.-Lines with the Greatest Mean Widening.

2. The Temperature of the Stars. By Sir NORMAN LOCKYER,
K.C.B., F.R.S.

3. Criteria of Stellar Temperatures. By H. F. NEWALL, F.R.S.

4. The Short-period Barometric See-saw and its Relation to Rainfall. By WILLIAM J. S. LOCKYER, M.A., Ph.D., F.R.A.S.

The first portion of this paper dealt with the short-period barometric see-saw which has been found by the author and Sir Norman Lockyer to exist between two antipodal regions on the earth's surface. This investigation, which has been published by the Royal Society,' has already been referred to in detail.

The essence of the second portion may be briefly summarised in the following

words:

The variation of the barometric pressure over India from year to year is the inverse of that over Cordoba, in South America-that is, when the pressure over

1 Roy. Soc. Proc., vol. lxx. p. 501, vol. lxxi. p. 135, and vol. lxxiii. p. 457.

India is in excess of the normal for a year or so, that over South America is deficient or below normal.

A study of this pressure variation for places widely scattered over the earth's surface shows that the earth's surface may be divided into two parts: one part behaving more or less like India, and the other like Cordoba. A classification of pressures of these two regions shows that a dividing line may be drawn round the earth, on the opposite sides of which barometric see-saws take place.

As it seemed possible that the knowledge of this regular barometric see-saw would render possible forecasts for approaching seasons, its relation to rainfall was investigated.

Since rainfall, generally speaking, accompanies low pressure, the inverted pressure curves were compared with the rainfall curves for several stations. The very close relationship between the rainfall and inverted pressure curves which was thus found to exist suggested that there was a possibility of forecasting wet and dry years.

The problem becomes more difficult the further the equator is left behind and the poles approached, but it is thought that when further investigation has been made the behaviour of the pressure and consequently rainfall variations in these regions will be more completely known.

5. The Relation between Solar Physics and Meteorology.
By Professor BIRKELAND.

6. Experiments with Kites in the Mediterranean.
By L. TEISSERenc de Bort.

7. The Relation between the Minima and following Maxima of Sun-spots. By ALFRED ANGOT.

Professor Wolfer published some time ago a revision and continuation of R. Wolf's table of relative numbers of sun-spots, so we have now, for every month or year, the mean intensity of sun-spots from 1749 to 1901.

When working at that table in order to verify or find some periodicities I met with a relation that was not mentioned, so I it presume may be new.

When the mean number of sun-spots during a year of minimum is under the average of all the minima, the value of the immediately following maximum is also under the average, and vice versâ,'

This will be apparent from the following table, which gives the relative numbers of sun-spots (r) during the years of minimum and the numbers (R) for the immediately following maxima, arranged according to the decreasing values of the minima. The numbers in brackets are the corresponding years: