^} .-> 1 •f 'V, \: v I \ *y 4 X. K Kj V I '\ xi X % K ^] :'^ V: V ^'* -4 4 lieprinted from the Transactions of The Astronojuical and rhysical Society of Toronto for 1S97. PERIODICITY OF MAGNETIC DISTUltBANCI'S. liY ARTHUR HARVEY, F.R.S.C. The Society is aware of the interest I have fult in Mr. Pursey's regular reports of his sokir observations, and that soon after lie com- menced them I began an enquiry into the periodicity of snn-spots ; not as to their maximum and minimum frequency, but as to whether, at more or less regular intervals, spots would break out in the same solar regions. To come to a conclusion on this subject, the first thing to be deter- mined is, of course, the rate of solar rotation, and, following my favourite method of m.iking an investigation for myself without reference to the studies of others, and of referring to authoi ities only after having gone as far as pcssil)]e independently, I set to work to ascertain what t he time was in which the Sun turned upon hisjixis. A stronomer s, as all menibers of tlie Society know, arrive at this rate of rotation from the passage of spots across the Sun's disc. INIost of my hearers will remember that Galileo was among the fir>t to observe these spots with the telescope. It was his " Dissertation on Hun-spots," which, published in 1()13, drew on him the censure and warning of the Cluuch, which declared it was contrary to Scripture to say the Sun moved. The words, '• and nevertheless it does move," attributed to that immortal physicist upon very doubtful authority need not be faintly muttered now, as it is said they were by him, and, curiously enougli, it is to an ecclesiastic, the Jesuit Scheiner, we owe the first fairly accurate deter- mination of the time it takes for the spots to traverse the disc. He stated in lG7o, after a long life devoted to observations, that the rotation period was 2o-33 days. It does not .seem to have occurred to the early astronomers to go beyond the simple question of the average time of sun-spot rotation, and successive observers occupied themselves with the cori-ect determination of that period only. Nor is this so easy as it may seem, for chatiges in form are always taking place, no spot having the .same contour for so much as a single day, some, moreover, having a proper motion of their own, which is more or less confusing. It has lately been recognized, too, that the rate of rby.atSfyiii.* va'j-ies 'in diticrert solar latitudes. Still, the idea of a nucleu^4i6."tBtJ.Slili,.'sun*durl€levl "bV ii^' atmosi)here, this .* • • . J ' i , , , : . > , ' ■i^ 57700 nucleus rotating at a steady rate while parts of the atmosphere have a Tlifferent rate, has not met with general acceptance, nor is it yet spoken of in any of tlie standard books I have consulted. Such, however, was the idea which, as you know, occurred to me, and if I can trace it to any other person's suggestion, it is to the late T. Sterry Hunt ; not, indeed, in connection with the Sun, but in that essay where he speaks of the dei)osition of the limestone strata on the Earth. He pictures our planet as having had at one time an atmos- phere charged with calcium vapour, as we know that of the Sun to be at present, and he tells of acid rains during the cooling process, to which our enormously thick deposits of carbonate of lime are due. I derive from him the conception of such a rain originating from condensation in the upper regions of our air and descending on the still intensely hot equatorial regions. Their heat would again vaporise that rain and disen- gage the fixed carbonic acid as gas. In its explosive ascent it would burst through the envelope of self-luminous cloud-laden air. An extra terrestrial observer might then have seen the bright vapours dispei'sed in those regions where the action was violent, and beheld a compara- tively dark spot, with irregular edges, changing like the spots we now see u])on the Sun. But the conditions upon the Sun cannot be the same as those upon the Earth were, even at its incandescent period. There cannot be seasons resembling those which from the earliest time in the Earth's liistory there must have been here. Some different cause for a dimin- ished temperature at the solar poles, if it exist, must be found. There- fore no argument can be strong which is based on such hypothetical analogy alone. I began then with a simple supposition based upon one sei'ies of observations, that uj)on the surface of the solar nucleus there are regions where chemical action is from some cause more violent than in others ; such action being sometimes forceful enough to dispel the upper photospheric cloud layer, and I proceeded to enquire how far other observed facts would justify this theory. Pardon me if I digress to define synodical and sidereal rotation. Imagine a line drawn from the centre of a rotating Sun through its equator to a Star. It would pass through a certain point on the solar surface, and, taking Scheiner's period, that point would in 25 '33 days have been turned around tlje solar axis ^nd .^i-guiu-. be on the imaginary line. That is tfie:tanIe«'o5 si3«i'eal iT3titiort.'**iiut*lf when we fir.st drew on the solar surface lave been turned around tlje solar axis ^nd ,?i,gaiv>.b£ ine. That is tfie:tjnIe«'o5 siawreal iT3titior{.'**iiuflf • • • '.' , ' ', ' ', ••• t ••••••• **• the line, the Eaitli wer"(B*p3Sling* across it, the point c "*! • • I • 1 !• • • ,♦ . .* ..*• • ••••/ .•••• •• central on the Sim's disc, as seen from the part of tlie Eai-th just on that line, would not be central when the point had again come round to its place with respect to the star, for the Earth would have gone a long way forward in its orbit. A longer period is therefore required for the given point on the Sun to become a second time central on his disc as viewed from the Earth. That difference, added to the sidereal rotation, gives us the synodical rotation of the Sun. Knowing the length of our year, the one is easily reducible to the other. Now, my first examination of Mr. Pursey's diagrams was of this nature. Taking any large spot, central on the disc on a given day, at what date was it again central by rotation ] Even though not traceable by rotation more than once or twice, then, assimiing the rate of rotation thus ascertained, would the spot reappear as central after quiescence ? Let me mention here one of the s])ots I thus followed. In this case I have ascertained from Mr. Maundei', the President of the British Astronomical Association, through Miss E. Brown, the President of its Solar Section, the heliographic latitude and longitude of the spot. Mr. Maunder writes that he once himself suggested the outbreak of spots in a " disturbed region " after temporary quiescence. He also states that this particular spot belonged to a " disturbed area." It seems a pity that with the facilities Greenwich Observatory possesses, the " sugges- 1896. d not have been thoroughly examined. Date of lentrality. Heliographic Lat. Long. Character of Spot. Intervals in Days. Feb. 24 Mar. 23 Apr. 20 —16' —ir —18^ 220° 215° 215° Fine Large Fine 28 28 June 14 July 12 — 18' — 14" 215° 217' Enormous Fine 55 i 28 Sept. o —18° 212° Group 56 Oct. .30 Nov. 26 Dec. 23 Jan. 21 Feb. 18 —IS'' —19° — ir 203° 148° 210° Noteworthy A spotted region Fine Fine 55 26 27 29 26 '7 + 7 -h 1897. But the rotation of apots is an imperfect means of determining that of the Sun proper, for, as Mr. Maunder puts it, " the synodic period varies for every latitude." As it is not possible to supi)Ose that the whole body of the Sun is contorted and that each frustum of the whole sphere revolves at a different rate, the spots must be connected with some kind of atmosphere, within which theie is a nucleus. At what rate does the 4 nucleus rotate, iuul how are we to get behind the veil of photospheric clouds to find the answer to this enquiry 1 It is a question of inqiort- ance, for if the spot s nie mere ly symptomatic of great disturbatices on the nurleus, caused by' volcanic eruptions from a reg'on in which intense chemical action is going on, and are seen whenever in such a region an outbreak occurs of sufficient violence to affect the solar atmosphere thiough all its layers, then, if we knew the exact rate of rotation of the nucleus, we could trace the history of spots for many long years — id entify _ with each other sj)Ots whose appearances are separated by considerable intervals of time — learn the relative activity of particular regions — and know where and perhaps wlien to look for great spots again. It also follows that the non-reappearance of sjjots over j)articular regions would show the phenomenon to be merely atmospheric and unconnected with the nucleus. As the breaking out of spots is said to cause the radiation of much heat, the temperature of the Earth must be hottest when the s})ot is turned towards us. The solar rotation period ought, therefore, to be noticeable by means of the theimometer. In examining the curves of temperature I found strong indications of a periodicity of between 27 and 28 days, but the irregularities were very baffling. Noticing next that the curves of temperature seemed to be very closely connected with magnetic curves, I thought I might find a safe guide in the latter, and began the enquiry, of which I have finished one stage only, which I shall not be able to complete in my lifetime, but which I have found .so deeply interesting that I shall have to continue it as long and as far as possible. A non-professional physicist need not fear to walk in such a path as this, for the maxima and mhiima of sun-spots — their increasing and diminishing frequency towards maxbnum and minimum, and the assertion of periodicity in these maxima and minima is due to Schwabe, a simple counsellor, of Dessau, who happened to have a 3-inch telescope, and for his amusement began solar observations which, continued for forty years, led him to announce these remarkable discoveries. I claim to be the first to endeavour to deduce from magnetic records the rate of rotation of the nucleus of the Sun. In the course of my enquiries I ha\ e found several whose paths of investigation approach and cross mine, but so far I have not found any who before me announced that magnetic disturbances could be made use of with abso- lute exactitude, for the determin.ition of this important point. I was almost frightened from my proposed couise by Maxwell. He dismisses terrestiiHl plienoineiia in one chapter of his work on magnetism. He says that a meniher of the Koyal Society of" fiondon found, after close examination, three different periods for magnetic disturbances ; one that of the rotation of the 8uu, and the otlier two tliose of the synodical and tropical revolutions of the Moon. The whole question, he adds, is full of intricacy and presents unusual difficulty. It, therefore, seemed prudent to attack the subject step by step. The first question to consider was the exact date to be us.siyi)ed to any giveii' magnetic disturbance. T ofTb'i*'1iere a" scale-diagram of the record of Horizontal Force for July, August and September of this year, where a dip begins on September 2nd and lasts for ten days. Which day shall we take for our reckoning ? I have adopted the method of taking the lowest marking. Moreover, this chart shows the mean read- ing for the day. I take the lowest hour of that day, but only do not fail to note any very low reading during the progress of the depression. In this case it is on the lltli, which is a repetend of the dips on July 17th and August 14th. I take the Toronto records alone, because disturbances are so synchronous all over the IfTorthern Hemisphere that it is only a multi- plication of tedious work to resort to others. In proof I give the com- parative mean daily curve for the observations at Toronto and at Tiflis (Caucasus, Russia), for the months of January, February and March, 1895. The difference in the gradients, which is but slight, and of no importance for my enquiry, would probably disappear if corrections were made for the different magnetisation of the respective bifilars. I take Horizontal Force, because I do not wish to resort to averages. A magnetic storm affects dip and declination too, but in somewhat different ways. 1 hope, as soon as the Toronto observations are printed, which they have not be m since 1848, fifty years ago, to examine the disturb- ances in these records too. Prof. Frank H. Bigelow, now of the Weather Bureau, Washington, D.C., whose work is near in character to mine, published a paper in the American Journal of Science for December, 1895, in which he proved that the curve of temperature followed that of magnetic disturb- ance very closely, and assigned 23.68 days as the period. He concluded that paper by saying that from all his tables he gathered the underlying suspicion that there was a nucleus to the Sun, which rotated at the same rate as the solar equatorial belt. At the recent meeting of the British () Association, Piof. liigelow showed me hin tables, and lie has favonied nie with a letter, in which ho says they are " the results of a very extensive computation of European ol)servations for the years 1878-1889, and the period can be seen readily by even a hasty glance at the sheets." Prof. Bigelow means by this tliat he has underlined the principal disturbed periods of each calendar month, and can readily see that the belt of disturbance thus indicated, which covers seveial days, occurs from three to four days earlier each month. I had seen this in all the tables, but found it much moie satisfactory to reduce them to graphic curves. By these I was confirmed in my view that this method, which I had considered before Prof. Bigelow's admirable paper was brought to my attention, was not sufliciently exact. The disturbed period of the month did indeed recur, but the disturbance was of different duration each month and its intensity was differently concentrated, a disturbance of one or two days' duration tending to lengthen and to divide into two parts, the following of which led to uncertainty and confusion. In proof I ask you to consider the Tiflis and Toronto curves just given. . . . The lines drawn perpendicularly across them answer to periods of .solar rotation, and you will easily trace the variation to which I allude, In Maxwell's work it is stated that most of the variations are due to terrestrial causes, but that some of the most violent distui-bances are believed to be connected with solar phenomena. The result of my study is to connect the disturbances chiefly, if not altogether, with the Sun, the ruler of his system in magnetic as in other particulars. There is doubtless a daily and an annual change in magnetic conditions, as well as a secular one, dependent on terrestrial causes, such as changes in temperature, but the great irregularities are all periodic, and if any of them are solar in their origin, so then are they all. The interferences of one disturbance with another are, as I believe, the causes of the seeming confusion. An outburst of activity on the northern hemisphere of the Sun will probably mask the effect of another on the southern hemispheie. An eruption on the side of the disc turned from us may interfere with the needle in a very different way from one on the side facing us. The Sun is always in a state of the most intense activity, many regions being disturbed at the same time. Facula? and protuberances, especially the latter, are, I think, as symptomatic of those solar conditions which produce magnetic effects as sun-spots are. It has puzzled many to disentangle the numerous inter- fering waves of pulsation, but I think I see how it is to be done. With your leave I .vill enter briefly into some details of tlie investigation. I coinniencod with a remarkable storm on January lOtli, ISiiA. There is a corresponding storm on February Ttli, twenty-eight days thereafter. Meantime another important disturbance occurred on January 19th, duly followed at an interval of twenty-eiglit days on February 16th. These two disturbances continue their rf'cord on the photographic traces throughout the year, and a third wave appears at the eiid of September. Of sixty-five periods for recurrence I can iden- tify fifty-one, and the average period is a little over 27] days. Please to note, on my charts, for your.S(!lves, those numerous recuiient tlips. (The charts were here exhibited and examined.) I have, however, given most study to the disturbance of 13th January, 1888. The mean of this day was the lowest of the month, and the reading at 10 o'clock was the lowest of the day as well as of the month. Having the key to the mystery furnished by the approximate period of 27| days, I traced this dip both forward and backward, and I present the following table of sixty-three periods for its occurrence : — 1 2 3 4 .I 6 7 8 9 10 n 12 13 14 15 16 17 18 19 2d 1887. C9 U CI p o Jan. Feb. Mar. April May June July Aug. Aug. Sept. Oct. Nov. Dec. 1888 Jan. Feb. Mar. April May May June o = a 4 S'i 3 0) ♦- o a o ax 24-4 20-10 19-16 1.5-2-.J 13-4 9-10 6-16 •2-. 2 30-4 •20-10 23-16 19-22 17-4 13-10 9-16 7-22 4-4 1-10 28 10 24-22 2 > Z O M -23-19 21- •20- 13-" ' ' NOTES AND HKMAUKS. The lii-st six observations show merely the i>reliniinnry inMicatintis of a no'eworthy disturbance, like the first niuttevlnKs of ii terrestrial storm. The series propei-ly begins with the depression of .liilv 7tli. Inter\als (.-alciilated from .(aniiary 13th, 1888, as a datum. 7-4 3-11 30-10 20 -22 -24-4 -2010 lV-10 13-10 ' V-ib' 1 10 The absolutely lowest reading of this small decided dip is on tht 2-2nd, at -2.h. Of small importance. First decided " pull" of a dip which lasts until •2-lth. Nothing noticeable. This and the Uth are days of mwlerate depression. Very slight dip. The lowest reading of the month. Again the lowest reading of the month. The lowest mean daily reading of the month. A great storm, the lowest reading of the year. Very perceirtiblo, but the violence of the periodical storm is iiioderatiiig. Still moderating, but the mean reading is still the lowest of the month. Still a considerable dip, though a temporai-y subsidence. The mean reading and the lowest reading also, ai-e the lowest of the month. Nothing shown in the mean reading. A decided dip at 11 hours. This is a decided dip ; lowest reading since 1st and until lOtli, though the ^th and Hth have lower means. The 4th and .oth are the lowest means of the month initil the 11th. This day has the lowest mean until Vth ; the reading is lowest until •20th. No indication of repeating. One should not forget that this is an autumn storm. (anniversary. Still no indications. It broke out in Julv of tlie last year. Wait for the s 1887. ■2.SC ."l^-i », t* c CD's cj a) III •0 '- .— • CO CS 01 ■S u th, a dav late. An important dipression, but increasingly late. An interference by an- other dist.nbaiicc probalile, which will retard this oue, while it con- tinues Storm seems to be widening and dividing. The reading is the lowest since the .'itli This day — one and a (piarter days before lime .'see No. 4! alsoi lias the lowest reading of the month, tbo\igh the meati of the I8th is lower. An im))ortant deju-i'ssion this day ai d ihe text. Also an imiioitant depression. I'isturlu d perinil lasts several days. A iioteworthv dip. Slight dip. [)e(-ided dijis Some interference on the 2' th. An important dip A decided depression Same lemark Obs. give a somewhat disturbed Icurve on (lib and 7th. the minimum for the month at 2'2 Nothing of moment to remark 'I'ifli A very slight depression. An important dip 'liMis Olis. ;. hours on -.'nd. The dip begins now and la-ts over 1st and 2nd Ai>ril. Tiflis minimum i8 Nothing of special inonunl, to record. |at lib on Ist. Same remark. The disturban(-e of th" previous week marks this, which should reai)pear in a moinh ' v two. Slight dei)ression and dist\n-bance. Storm of the week before is leas violent than during the .May term tliereof. This is the most inioortant depression of the month, the mean reading being the lowest, also the reading at the hour named. The depression seems to have been neutralised by the remains of that which preceded it in ril and .May. That was a spring storm and shoidd not longer affect this. The most imi>ortant disturbance of the month and the lowest dip. A very low reading. So, t 10, is the reading at Tiflis on the 9th at Ifh., which mav be the date «c should regard n.ost. 9 Here the outbreak ends as an important storm. On these sixty- three dates, eight give the lowest reading of the respective months, and S^Btl Others are very close to it. Only fourteen show nothing noticeable, so that in forty-nine cases indications of repetition are traceable. If we reject the first mutterings of the storm and commence with the impor- tant and clearly recurrent disturbance of July, 7 days 4 hours, 1887, we have 57 occurrences, or 5G periodical opportunities for retuiJi in the 1525 days 23 hours between that time and September 10 days 3 hours, 1891, which gives us 27 "249 days for the period. Going back into pre- photographic times, we can connect this storm with that of 8th October, 1886, prefaced by an equally severe storm of 10th Septeml)er. We see it on 29th ^larch, 5th February, 9th January of that year, and, by following the .series of principal outbreaks still further back, 1 can identify it with the dip of January, 15 days 6 hours, 1841 — the first year of our Toronto observations — which was the most important of its month, and recurred 2nd and 29th June, 23rd to 25th July, 15th October, 11th November and 8th December of that initial year. From that beginning to the end of the series as given in my table, there are 679 rotations in 18'499 days 21 hours, which gives 27'24575. . . . days as tiie average interval between tlie magnetic disturbances which aiise from this more or less i)ermanently active area on the inner body of the Sun as it is turned around into tlie position in which it most strongly affects the magnetic conditions of the terrestrial field. In other words, this is the synodical rotation period of the Sun, where sidereal rotation, therefore, occu[»ies 2535447 days or 25 days, 8 hours, 30 minutes, 26'2 seconds. I cross with a line, at the proper intervals, the curves drawn from the daily magnetic observations for all the intervening and subsequent years, and we are thus enabled to follow with ease the history of the storm. After September, 1891, it was interfered with by other disturbances, notably by one of November 20th, which itself recurred raaikedly on January 29th, 1892, still more violently on May 18th, and broke the magnetic record on July 13th and 14th. During suppression, our storm appears by faint traces to be struggling to reassert itself, and on February 25th, 1894, it seems to coalesce with another eruption, forming a grand series of magnetic waves. It reappears on January 19th, 1895, repeatiiig vigor- ously on February 16th, March 14th and 15th, April 11th and 12th, but only to subside into insignificance again, which continued to the time I prepared a paper for the British Association. I said in that 10 paper that it might at any time be renewed, and on July 31st it broke out again ; it is in evidence also on the 23rd Sejjtember, In discussing my paper, I said we might look for traces of it on August 27th, September 23rd, October 20th and 21st, etc., and this was reported in the city papers. Tlie present is a very quiet magnetic year, and great disturbances are not to be expected, but on August 27th we had a low reading (33-2) at 21 o'clock, with a still lower one (32-4) on the 28th at 10 o'clock, while on September 23rd at 10 o'clock, we had the absolutely lowest reading since recovery from a preceding dip on the 11th. I think thi.s is the first attempt at predicting a magnetic phenomenon, and it was successiuL* '^"^^v^lheCentury magazine, for the present October, in a very inter- esting article on auroras, the following paragraph occurs : — " In 1872, Prof. Young noticed a disturbance in the chromosphere in the neighbourhood of a sun-spot, and upon asking the astronomers at Greenwich and Stonyhurst to examine their magnetic records, it was found that great disturbance had occurred about that time. Ten years later, the astronomer at Greenwich sent out a message that read some- thing like this : ' Remarkable sun-spot now visible .... area of whole spot 247 — 100,000 of the Sun's visible surface.' Try to imagine what this means, and fancy youi',self on the Sun while that tremendous storm was in progress. We know that here on Earth there was a magnetic storm with auroral displays that beggar description. Beginning a little before daylight on November 17th, 1882, not a wire of the Western Union Telegraph Company could be used for three hours. * * It so happened that about the pole, that year, were clustered representatives from twelve nations. * * November 14th to 19th, 1882, was a period never to be forgotten by these Arctic prisoners. While we at home saw the display of a decade, the observers of the frozen north * * saw visions glorious by day as well as by night, and felt, perhaps, some recompense for their isolation and peril." I have brought my diagrams to this meeting that you may see the identification of this storm with that of which we have been following the history, and see for yourselves the long continued violence of it, as * In October the disturbance began towards the close of the 18th, a day and a- half before time, but lasted until the 20th. In November an important disturb- ance covered the 16th, 17th and 18th, the lowest reading of the month being on the 17th, at 9h. 53m. The December disturbance has not at this writing been investigated. — Author's Note. 11 well as the intensity which at one time drove the recording pen clean oft' the paper. The Century article contains the following illustrations ; — 1. West end of an auroral band, photographed February 1st, 1892. My curves show a noticeable dip on February 2nd. 2. Photograph of sun-spot, August 8th, 1893. These are important spots. The article does not say they were central at that date. Here, however, is an important magnetic storm, August Gth and 7th, 1893. 3. Aurora of December 21st, 1882. Observe the important magnetic dip at that date. Sun-spots cannot yet be said to be coincident, either on first ajipear- ance or when centrality occuis, with magnetic outbursts, nor can we yet say that all magnetic storms cause auroras. I simply show tlie magnetic curves, along with the Century's article and illustrations, to illustrate the advantage and convenience of having them in this shape and to indicate the possibility of a future co-ordination of these allied phenomena. In Nansen's Farthest North eleven auroi'se are specially referred to. Two are co-incident with remarkable magnetic dips, three with slighter ones. The rest were evidently strictly local phenomena, and produced no effect on the daily means. The chief difficulty I have thus far encountered in establishing to my satisfaction the theory of permanently active areas on tiie Sun is an unexpected one. You are aware that, having determined the rate of solar lotation, I mapj)ed on a chart of the Sun the localities of the pro- tuberances recorded by the Italian spectroscopists. This I did for each of the years 1894, 1895 and 1890. The results I have heretofore shown you ; the charts resemble the belts of Jupiter, havirg marked areas of disturbance and of quiescence. But in one year, the polar belts are further north and south than in the next. The inference is primd facie adverse to my contention. Much more has to be done, then, before the theory can be established. IMeantime it is something to have arrived at a definite idea of the rotation of the Sun's nucleus, from magnetic data. I am indebted to Prof. Bigelow for a reference to soine other calcu- lations of the Sun's rotation period from terrestrial phenomena, given in the Bepertorium des Physik iov 1866. We have no copy of the work 12 in Toronto, and I cannot say what means have been used, but the table given me by Prof Bigelow is as follows, our two figures being added to those in the liepertoriiwi : — Deductions of S un's Svnodic Rotation from Terrestrial From Magnetic do. From Sun-spots. Phenomena. m ,- J^^^' *. 26-68....Bigelow. 27-51 27 05 25-87 27 2457..Harv3y. 27-32 26G9 25 -86 27-27 2C39 2'>-32 25-86 25-83 ^l7'th" "• 26 24 25-82 27-10 26 05 25-79 26-85 26-03 25-66 25 96 25-47 25-92 The mean of all these is 2G-43, but Prof. Bigelow very justly says that " this is a question not to be settled by a majority.'' He remarks on the curious fact that the results from terrestrial phenomena tend persistently to be much lower than those from sun-spots and other visible solar phenomena. This is, however, not the case with my results, which agree closely with them. So also do tho.se of Dr. M. A. Veeder, of Lyons, N.Y,, a corresponding member of ours, who has tabulated auroral displays, and finds that a period of 27|- days will fit the principal recorded occurrences of note- worthy brilliancy. I ought not to conclude this paper without referring to the work of Mr. Carlos Honore, of Montevideo. By investigating the late Prof. Gould's thermometrical observations at Cordoba, Argentine Eepublic, he finds that there is a periodicity of hot and cold waves which he places at 27-241326 daj s. He has published his views in a work called the Law of Solar Radiation (Loi du Rayonnement Solaire, Montevideo, 1896). He adopts as the commencement of his periods January 1st, 1894, at mean noon, in the longitude of the Villa Colon Observatory. If I were rearranging my work, I should take the 6th of March and the 6th September as the dates for commencement, or the 4th of June and the 6th of December, the dates at which the Earth's orbit crosses the plane of the solar equator. Thus I should divide the tables and the 13 correspon(1ing curves into periods influenced chiefly by disturbances on the noithern and southern solar hemispheres, respectively. Mr. Honore considers that solar vibration proceeds on lines parallel to the phine of the solar equator, and thus explains the disappearance after several rotations of the maxima and minima of magnetic distur- bances at the calculated periods. Prof. Bigelovv has noted what he calls the inversion of the solar magnetic curve, and thinks magnetic force proceeds from the Sun in great curves, from one of his poles to the other, like those of an ordinary spherical magnetic field here, which Mr. J. K. Collins has shown this Society by means of photographs of iron filings. These curves embrace the Earth, nay every planet in the system. I have not gone further than to ob.serve that the principal disturbances of the early months of our year are not identical with those of the fall months ; their period is the same, but the periods of one are not continued with anything like the same intensity throughout our year. Many things will have to be explained before I can give credence to the view that magnetic currents can be emitted from an intensely hot sj)here, and I am forced to favour a theory of some form of radiation or vibration which acts inductively and indirectly upon this and other planets. Mr. Honore coincides with me in referiing his distuibance period to the effect of the rotation of the Sun's nucleus, but both he and Prof. Bigelow have taken a much shorter range of observations than I have. To this I attribute the slight difference between Mr. Honord's period and mine. Prof. Bigelow, however, differs so much from both that either his method is faulty or those of Mr. Honore and myself are Lastly, I wish to express my obligation to Mr. R. F. Stupart, the Director of the Magnetic and Meteorological Observatoiy here, for allowing me free access to the records, and for many kind answers to enquiries. t • • • • < • ••• • ..• ... .•••;::