the Birmingham and Midland Institute-is to be congratulated on the good work he has done in this connection. The book is illustrated with 72 figures, which agree with the simplicity and clearness of the diction, and questions are found at the end of each chapter, which have been well prepared to test the learner's apprehension of its contents. We are pleased to be able to recommend this little work, as a foundation for the study of the metallurgy of iron and steel. OUR BOOK SHELF. On the Creation and Physical Structure of the Earth. By J. T. Harrison, F.G.S., M. Inst.C.E. (London: Longmans, 1889.) THIS book brings to mind one of the most winning of the vagaries of childhood. A bright child of an inquiring turn will sometimes sit with comical sedateness listening to the talk of its elders. It may afterwards be overheard repeating to one of its playmates, or to some lucky adult who has the knack of winning its confidence, such detached scraps of the conversation as have found a resting-place in its little brain; and, conscious even at its early age of the necessity of some continuity in a narrative, filling up the gaps with inventions or criticisms of its own, charming every way, but mainly on account of their utter want of connection with the subject of the conversation which it is attempting to report. So our author has listened to the teaching of many geologists, and has culled many detached passages from their writings: these he repeats to the world in a book, printing between them comments and lucubrations of his own, about as innocent and as little apposite as the child's prattle-hardly so amusing, however. The following passage is a fair sample of the writer's own share in the book. The termination of the Secondary Period, which introduced these altered conditions of the surface of the northern hemisphere, was really the commencement of what is called the Glacial epoch in Europe. We have noted signs of glaciation during the deposition of the upper chalk in India and North America, but now the conditions which induced that glaciation are extended in such a manner as to unite these districts, and produce that enormous accumulation of snow and ice at the North Pole, the weight of which in the Miocene epoch depressed the crust in that region and upheaved the mighty mountain ranges to which I have just referred." The book bristles with cataclysms and catastrophes. The shifting of a thin crust on an internal nucleus which it does not fit, and incessant protrusions of granite, are invoked to account for phenomena which every-day people still persist in thinking are satisfactorily explained by every-day causes. But the author is one born out of due time-two centuries too late. How he and Burnet would have enjoyed a crack together! But there is this to be said, the "Sacred Theory of the Earth" is Burnet's own the staple of the present work consists of extracts from the works of others. The mottoes are verses from the first chapter of Genesis, but their relevancy to the subject-matter of the chapters which they head is not obvious. A. H. G. Through Atolls and Islands in the Great South Sea. By F. J. Moss. (London: Sampson Low, 1889.) MR. MOSS-a member of the House of Representatives, New Zealand-started from Auckland, in September 1886, in the schooner Buster, for a voyage among the islands and islets of "the outer lagoon world." He was absent seven months, and during that period he crossed the equator six times, and visited more than forty islands among the least frequented groups. In the present volume he sums up the impressions produced upon him by what he saw and heard in the course of his voyage. Mr. Moss, in dealing with matters which really interest him, shows that he is an accurate observer and a man of sound judgment. His style, although plain and unpretending, is well fitted for the task he has fulfilled. The best parts of the book are those in which he tries to convey some idea of the daily life led by those natives whose customs he had an opportunity of studying. He appreciates warmly some aspects of the various Polynesian types of character, but thinks that the people are likely to degenerate rapidly, unless they can be provided with a better class of native teachers than most of those to whom the duty of guiding them is now intrusted. What is needed, he thinks, is, that the islanders shall have in their work and in their amusements freer scope for the imaginative powers with which they are endowed, and the exercise of which is too often foolishly discouraged. Everything Mr. Moss has to say on this subject deserves the serious consideration of those to whom his warnings and counsels are either directly or indirectly addressed. LETTERS TO THE EDITOR. [The Editor does not hold himself responsible for opinions expressed by his correspondents. Neither can he undertake to return, or to correspond with the writers of, rejected manuscripts intended for this or any other part of NATURE, No notice is taken of anonymous communications.] Who Discovered the Teeth in Ornithorhynchus ? IN NATURE of November 14 (p. 31), Profs. Flower and Latter criticise my note which appeared the week previous (November thorhynchus. They promptly dismiss my claim that Sir Everard 7. p. 11), concerning the discovery of teeth in the young OrniHome discovered the teeth of the young Ornithorhynchus, by stating that the structures described and figured by Sir Everard are the well-known cornules of the adult animal. If they will take the trouble to turn to the plate cited by menamely, Plate lix. of the second volume of Home's " Lectures," 1814-and will read the accompanying explanation, they will see that Home was familiar with the teeth of both the young and the old animal. 66 For the benefit of those who may not have access to Home's "Lectures,' I here reproduce outline tracings of two of his figures. Plate lix. Fig. 2, shows the teeth of the young Ornithorhynchus the "first set," as Home says, to show that there are two grinding teeth on each side." The next figure is a similar tracing from the succeeding plate in Home's "Lectures" (Plate lx.), which represents, to again use Home's words, "the under jaw of the full-grown Ornithorhynchus paradoxus, to show that there is only one grinder on each side." Both of these figures are natural size. In the face of these facts, further comment seems unnecessary. I admit, of course, that Home did not discover the chemical composition of the teeth of the young animal-this was Poulton's discovery. Washington, D.C., November 30. C. HART MERRIAM. [We do not reproduce the outlines sent, as anyone interested in the subject may see the originals, not only in Home's "Comparative Anatomy," but in the Philosophical Transactions, where they first appeared.-ED. NATURE.] I SHOULD be very sorry to deny the credit of any discovery to Sir Everard Home, or anyone else, if any evidence could be shown of its having been made. Of the figures cited by Dr. Hart Merriam, that of the younger animal seems (as far as can be judged from the roughly executed engraving, with the assistance of the descriptive text) to represent the horny plates, showing the hollows from which the true teeth have recently fallen; that of the old specimen, the same plates after they are fully grown, and their surfaces worn down by attrition. This difference led Home to conjecture that these plates were changed during the growth of the animal-a view which was corrected by Owen ("Comp. Anat. of Vertebrates," vol. iii. p. 272), by the statement " that "each division or tubercle of the [horny] molar is separately developed, and they become confluent in the course of growth.' By the way, no one can have been better acquainted with the work of Home than his successor in the Hunterian Chair, Sir Richard Owen; and yet, in his numerous references to this subject (Art. Monotremata," Cyclop. Anat. and Physiology"; "Odontography"; "Comp. Anat. of Vertebrates," &c.), no trace is shown of any knowledge of a discovery which could not have failed to have interested him, if it had been made before his time. 66 If a cursory perusal of Sir Everard Home's first account of the mouth of the Ornithorhynchus (in the Philosophical Transactions for 1800), or any interpretation placed upon his figures, might lead anyone to infer, with Dr. Merriam, that the real teeth of the young animal had been discovered at that time, the best possible authority may be conclusively cited against such an idea, no other than that of Home himself, who, in his later description of the same specimen ("Lectures on Comparative Anatomy," 1814), describes the organs in question as "the first set of cuticular teeth"-an expression quite incompatible with their being the teeth described by Mr. Poulton and Mr. Oldfield Thomas. It really seems superfluous to have to remind a zoologist of such high repute as Dr. Hart Merriam that the difference between teeth with the structure and mode of growth which characterize these organs in the Mammalia generally, and the horny epithelial plates of Ornithorhynchus, is not merely one of "chemical composition.' W. H. FLOWER. The Pigment of the Touraco and the Tree Porcupine. ATTENTION has been lately again directed to the red pigment in the wing feathers of the touraco, which has been stated by several observers to be soluble in pure water. Prof. Church, who was the first to experiment upon this pigment (The Student, vol. i., 1868; Phil. Trans., 1869), quotes Mr. Tegetmeier and others, to the effect that this pigment can be washed out of the feathers by water. Later, M. Verreaux (Proc. Zool. Soc., 1871) confirmed these statements from his own experiments while travelling in South Africa; attempting to catch one of these birds whose feathers were sodden with rain, he found that the colour stained his hands "blood-red." A few years ago Prof. Krukenberg ("Vergl. Phys. Studien ") took up the study of turacin-as Prof. Church termed the pigment-and added some details of importance to Prof. Church's account; Krukenberg, however, contradicted certain of the statements quoted by Church with reference to the solubility of turacin in pure water, remarking that the pigment in the dead bird is insoluble in water. writer in the Standard of October 17 is able "partially to confirm" the assertion that turacin is soluble in pure water. Seeing that there is some conflict of opinion with regard to this matter, I think it worth while to state that I found it quite easy to extract with tap water (warm) some of the pigment from a spiritpreserved specimen of the bird; only a very small amount couid be extracted in this way, and the feathers were not perceptibly decolorized even after remaining in the water for a fortnight. I also experimented upon a feather just shed from one of the specimens now in the Zoological Society's Gardens; this was steeped in water for some time without any effect being visible, but after a period of two days the water became stained a very faint pink. The touraco, however, is not a unique instance of a terrestrial animal with an external colouring matter soluble in water. I am not aware whether other cases have been recorded, but I find a pigment of a similar kind in a South American tree porcupine (Sphingurus villosus). A This porcupine has bright yellow spines which are for the most part concealed by abundant long hair. The spines them. selves are parti-coloured, the greater part being tinged with a vivid yellow; the tip is blackish-brown. I was unable to extract this pigment with chloroform, or with absolute alcohol even when heated; like so many other colouring substances which are insoluble in these fluids, the pigment could be extracted by potash or ammonia; I found also that tap water, warm or cold, dissolved out the yellow colour; the action was slower than when the water was first rendered alkaline by the addition of ammonia, but, unlike the touraco, the pigment was nearly, if not quite, as completely dissolved. The skin, from which the spines were taken, was a dried skin of an animal recently living in the Zoological Society's Gardens; it had not been preserved in alcohol or treated in any way which might lead to the supposition that the pigment was chemically altered. There is, therefore, a considerable probability that in the living animal the pigment is also soluble in water. I believe that this yellow pigment is undescribed, but I have not yet completed my study of it; in any case, it is not zoofulvin or picifulvin, or any "lipochrome." FRANK E. BEDDARD. Exact Thermometry. IN the account which Prof. Mills has given (NATURE, Decem ber 5, p. 100) of M. Guillaume's "Traité pratique de la Thermométrie de précision," the permanent ascent of the zero-point of a mercurial thermometer, after prolonged heating to a high tem perature, is stated to be due to compression of the bulb-rendered more plastic by the high temperature-by the external atmospheric pressure. The constant slow rise of the zero-point of a thermometer at the ordinary temperature is mentioned by Prof. Mills; and the late Dr. Joule's observation of this change in a thermometer during twenty-seven years is specially alluded to. It may, I imagine, be taken for granted that after the lapse of a sufficient length of time-possibly many centuries-a final state of equili brium would be attained; and it has always appeared to me that the effect of heating the thermometer to a high temperature is simply to increase the rate at which this final state is approached. It is my impression that, owing to the more rapid cooling of the outer parts of the bulb after it has been blown, the inner parts are in a state of tension, as, to a very exaggerated degree, in the Prince Rupert's drops; and that it is the gradual equalization of the tension throughout the glass that causes the contraction; in other words, that the process is one of slow annealing. This explanation appears to be supported by the facts—(1) that when a thermometer is exposed for a long time to a high temperature, the zero-point rises rapidly at first, then more and more slowly, and finally becomes constant or nearly so; (2) that the higher the temperature the more rapidly is this state of equilibrium attained. I do not know of any experimental evidence that the rate of ascent is influenced by changes of external pressure, and it seemed to be desirable to test the point. In order to do this I have exposed three thermometers, A, B, and C, constructed by the same maker and of the same kind of glass, to a temperature of about 280° for several days in the same vapour-bath, under the following conditions:-The thermometers were all placed in glass tubes closed at the bottom (C being suspended from above), and the tubes were heated by the vapour of boiling bromonaphthalene. One of the tubes-that containing thermometer C-was exhausted so as to reduce the the air. external pressure on the bulb to zero; the others were open to In thermometer A there was a vacuum over the mercury, but air was admitted into B and C to increase the internal pressure. Consequently, the bulb of A was exposed to a resultant external pressure equal to the difference between the barometric pressure and that of the column of mercury in the stem of the thermometer; the internal and external pressures on the bulb of B were approximately equal; lastly, the internal pressure on the bulb of C was the sum of the pressures of the column of mercury in the stem and of the air above it, while the external pressure was zero. The following results were obtained :— 1'15 I'20 Total rise of zero-point... The thermometers were heated until 5 p.m. each day, and the zero-points read on the following morning. If the diminution of volume of the thermometer bulb, usually observed, were due to external pressure, the zero-point of A should have risen, that of B should have remained nearly stationary, while that of C should have fallen. Instead of this, however, the zero-points of all three thermometers rose at nearly the same rate; therefore the yielding of the bulbs to pressure, owing to the plasticity of the glass, if it occurred at all, had no sensible effect on the result. SYDNEY YOUNG. University College, Bristol, December 12. Locusts in the Red Sea. A GREAT flight of locusts passed over the s.s. Golconda on November 25, 1889, when she was off the Great Hanish Islands in the Red Sea, in lat. 13° 56 N., and long. 42° 30 E. The particulars of the flight may be worthy of record. It was first seen crossing the sun's disk at about II a.m. as a dense white flocculent mass, travelling towards the north-east at about the rate of twelve miles an hour. It was observed at noon by the officer on watch as passing the sun in the same state of density and with equal speed, and so continued till after 2 p.m. The flight took place at so high an altitude that it was only visible when the locusts were between the eye of the observer and the sun; but the flight must have continued a long time after 2 p.m., as numerous stragglers fell on board the ship as late as 6 p.m. The course of flight was across the bow of the ship, which at the time was directed about 17° west of north, and the flight was evidently directed from the African to the Arabian shore of the Red Sea. The steamship was travelling at the rate of thirteen miles an hour, and, supposing the host of insects to have taken only four hours in passing, it must have been about 2000 square miles in extent. Some of us on board amused ourselves with the calculation that, if the length and breadth of the swarm were forty-eight miles, its thickness half a mile, its density 144 locusts to a cubic foot, and the weight of each locust of an ounce, then it would have covered an area of 2304 square miles; the number of insects would have been 24,420 billions; the weight of the mass 42,580, millions of tons; and our good ship of 6000 tons burden would have had to make 7,000,000 voyages to carry this great host of locusts, even if packed together III times more closely than they were flying. Mr. J. Wilson, the chief officer of the Golconda, permits me to say that he quite agrees with me in the statement of the facts given above. He also states that on the following morning another flight was seen going in the same north-easterly direction from 4.15 a.m. to 5 a.m. It was apparently a stronger brood and more closely packed, and appeared like a heavy black cloud on the horizon. The locusts were of a red colour, were about 2 inches long, and of an ounce in weight. G. T. CARRUTHERS. A Marine Millipede. IT may interest "D. W. T." (NATURE, December 5, p. 104) to know that Geophilus maritimus is found under stones and sea-weeds on the shore at or near Plymouth, and recorded in my "Fauna of Devon," Section "Myriopoda," &c., 1874, published in the Transactions of the Devonshire Association for the Advancement of Literature, Science, and Art, 1874. This species was not known to Mr. Newport when his monograph was written (Linn. Trans., vol. xix., 1845). Dr. Leach has given a very good figure of this species in the Zoological Miscellany, vol. iii. pl. 140, Figs. 1 and 2, and says: "Habitat in Britannia inter scopulos ad littora maris vulgatissime." But, so far as my observations go, I should say it is a rare species. See Zoologist, 1866, p. 7, for further observations on this EDWARD PArfitt. animal. Exeter, December 9, 1889. Proof of the Parallelogram of Forces. THE objection to Duchayla's proof of the "parallelogram of forces is, I suppose, admitted by all mathematicians. Το base the fundamental principle of the equilibrium of a particle on the "transmissibility of force," and thus to introduce the conception of a rigid body, is certainly the reverse of logical procedure. The substitute for this proof which finds most favour with modern writers is, of course, that depending on the 'parallelogram of accelerations." But this is open to almost as serious objections as the other. For it introduces kinetic ideas which are really nowhere again used in statics. should therefore propose the following proof, which depends on very elementary geometrical propositions. The general order of argument resembles that of Laplace. I I adopt the "triangular" instead of the "parallelogrammic " form. Thus, if PQ, QR represent in length and direction any directed magnitudes whatever, and, if these have a single equivalent, that single equivalent will be represented by PR. OP = OQ, QP = ON, NQ, QR, RP = OM, MP. (4) Finally, by (1), theorem holds for isosceles right-angled triangle; .. by (2) it holds for right-angled triangle containing angle 90°÷2"; .. by (3) it holds for right-angled triangle containing angle m. 90°÷2": i.e. for any angle (as may be shown, if considered necessary, by the method for incommensurables in Duchayla's proof). = Hence, if AD be perpendicular on BC in any triangle, Llandaff House, Cambridge, November 12. Glories. MR. JAMES MCCONNEL asks in NATURE (vol. xl. p. 594) for acccounts of the colours and angular dimensions of glories. I saw a good instance of the phenomenon on Lake Superior, June 17, 1888, and, having had my attention called to the value of accurate descriptions in such cases by Mr. Henry Sharpe's "Brocken Spectres," I examined it carefully. The shadow of my head on the mist was surrounded by a brilliant halo or glory, slaty-white around the head, followed by orange and red; then a circle of blue, green, and red, and the same colours repeated more faintly. The diameter of the innermost and brightest circle of red, as measured on the graduated semicircle of a clinometer, was 4°. There was also a very distinct, but nearly white, fog-bow outside, of 42° radius, as measured in the same way. A. P. COLEMAN. Faraday Hall, Victoria University, Cobourg, Ontario. THE appointment of a Commission at the present time to investigate the action of chloroform as an anæsthetic might to many seem an anomaly. For the use of chloroform as an anesthetic was introduced over forty years ago: it was in November, 1847, that Prof. Simpson, of Edinburgh, first brought this valuable agent before the medical profession. Since that time, the use of chloroform has enormously extended, especially in our country, and although there are other valuable agents of the same class-such as ether and nitrous-oxide gasyet there is a universality of opinion that the employment of chloroform has in many cases a special advantage. Considering the extensive use of the agent, and the progress which has been made of late years in the study of the action of drugs in man, it certainly is surprising that the knowledge of the effect of chloroform on the different parts and organs of the body is not complete. This is not altogether from want of attention to the subject; because, previous to the Hyderabad Commission, at least two Commissions were appointed with the view of investigating the action of chloroform and its occasional serious effects. These Commissions were appointed by the Royal Medical and Chirurgical Society of London, and by the British Medical Association, and they were composed of men who, from their knowledge of experiment and acquaintance with practical medicine, were competent to discuss the question. The two Commissions arrived at the same conclusions as the distinguished French man of science, Claude Bernard, had published years before, and these conclusions tallied with the teaching of the great London medical schools. Chloroform and other anæsthetic agents have a peculiar position they are powerful drugs used, not for disease itself, but for the purpose of allowing an operation to be performed, preventing the pain which would otherwise be felt, and relaxing the contraction and spasms of the muscles, so that the surgeon can more readily and accu rately operate. The administration of the anesthetic is something, then, outside the diseased condition; so that its use ought theoretically to be perfectly harmless to the sick person. Unfortunately it is not always so, and deaths from chloroform are, although rare, by no means unknown. The administrator of chloroform is therefore a person of great responsibility: he has to watch carefully the effect of the agent on the patient, to notice any unfavourable change that occurs, and to adopt measures to counteract any bad effects which appear. The knowledge of the mode in which chloroform causes danger to the life of the patient is therefore of vast importance; for, if the administrator knows the signs of danger, there is more likelihood of counteracting a fatal result. These fatal results, which are among the saddest that occur in medical practice, ought, if possible, to be avoided. What, then, is the danger to life of chloroform? Or, to speak more fully, what particular part of the body does chloroform injuriously affect when there is danger? This is just the point that the various Commissions have attempted to settle. In the Scotch schools, more especially that of Edinburgh, it has been taught that the great danger of chloroform was in failure of respiration; meaning by this that the danger-signal of chloroform was the stoppage or irregularity of the breathing. As a corollary to this belief, it was considered that the heart was only affected after the breathing had become interfered with; that, in fact, the respiration stopping, the blood was not oxygenated, so the heart stopped beating. This was the teaching of the great Edinburgh surgeon, Syme. The English (and especially the London) teaching was almost directly opposed to this. It was taught, and is still taught in the London schools, that the great danger from chloroform arose from its effect on the heart, which stopped beating before the respiration ceased. Which, then, of these two doctrines is true, or are both true? The decision of this question is, as we have stated, one of vast importance; but it must be remembered that, whichever is right, the administrator of anæsthetics always pays attention to both the beating of the heart and the regularity of the respiration. Surgeon-Major Lawrie, one of the prominent members of the Hyderabad Chloroform Commission, says that "it is possible to avert all risk to the heart by devoting the entire attention to the respiration during chloroform administration." Medical opinion in England, both of that of experts (professional anesthetists) and of the general profession, is distinctly opposed to this view; and the administrator who does not attend to the pulse, as well as to the breathing, is certainly neglecting one of the main paths by which Nature shows us what is going on inside the organism. From the statement of Surgeon-Major Lawrie just quoted, it will be seen that the Hyderabad Chloroform Commission came to the conclusion that the danger from the administration arose, not from the heart, but from the respiration. This view was strongly combated in our contemporary, the Lancet. The importance of the question led the Nizam of Hyderabad to obtain the services of a scientific medical man from England to go out to India and attempt to settle the question. Dr. Lauder Brunton, F.R.S., consented to go; and, well known as he is for his life-long devotion to the experimental investigation of the action of remedies and their practical application, it was considered probable that his aid in the research would lead to interesting and important results. From the somewhat scanty news of the results which have been telegraphed to England, it seems likely that the investigation now progressing at Hyderabad will tend to revolutionize existing views as to the action of chloroform. Dr. Brunton's views as regards the dangers of chloroform before he left England were clearly expressed in his well-known" Text-book of Pharmacology." In it he says that "the dangers resulting from the employment of chloroform are death by stoppage of respiration and death by stoppage of the heart;" he lays as much stress on the effect on the heart as on the respiration, and he proceeds to affirm that too strong chloroform vapour may very quickly paralyze the heart. This view is, indeed, similar to the one we have already mentioned as taught in the London schools of medicine. It is also well known that death may occur soon after chloroform has begun to be administered, from the heart being affected. If the operation is begun too soon, fainting from pain may supervene, and a fatal result occur: this has always been strongly insisted upon by Dr. Brunton. Surgeon-Major Lawrie says that in such cases it is not the chloroform that acts on the heart, but simply that there is fatal syncope or fainting. From the large number of experiments on animals which Dr. Brunton has performed in India, in conjunction with the Hyderabad Commission and a medical delegate of the Indian Government, it appears that the "danger from chloroform is asphyxia or an overdose;" there is none whatever from the heart direct. This statement is a distinct reversal of the view generally held in England. It means that chloroform causes a fatal result by affecting the respiration or by too much being taken into the system and affecting the brain; and that there is no direct paralysis of the heart from the chloroform. A perfectly impartial opinion cannot, however, be formed from the scanty records of the investigation which have been as yet received in England. We must wait for fuller details of the experiments before a final judgment can be passed. It is well, however, to point out that the prevailing view in England has been founded, not only on experiments on the lower animals, but also on the extended clinical observation of two generations of medical men. Clinical observation is not so accurate or so lucid as that of direct experiment, but it has its value, and one by no means to be despised in a case where it is so extensive, and directed to a subject of such great importance, not only to the medical profession, but to the general public, as the question of the administration of chloroform. IN ON THE CAVENDISH EXPERIMENT. N the last number of the Proceedings of the Royal Society (vol. xlvi. p. 253), I have given an account of the improvements that I have made in the apparatus of Cavendish for measuring the constant of gravitation. As the principles and some of the details there set out apply very generally to other experiments where extremely minute forces have to be measured, it is possible that an abstract of this paper may be of sufficient interest to find a place in the columns of NATURE. In the original experiment of Cavendish (Phil. Trans., 1798, p. 469), as is well known, a pair of small masses, mm (Fig. 1), carried at the two ends of a very long but light torsion rod, are attracted towards a pair of large masses, M M, thus deflecting the arm until the torsion of the suspending wire gives rise to a moment equal to that due to the attraction. The large masses are then placed on the other side of the small ones, as shown by the dotted circles, and the new position of rest of the torsion arm is determined. Half the angle between the two positions of rest is the deflection produced by the attracting masses. The actual force which must be applied to the balls to produce this deflection, can be directly determined in dynamical units when the period of oscillation and the dimensions and masses of the moving parts are known. In the original experiment of Cavendish, the arm is 6 feet long, the little masses are balls of lead 2 inches in diameter, and large ones are lead balls I foot in diameter. Since the attraction of the whole earth on the smaller balls only produces their weight, i.e. the force with which they are attracted downwards, it is evident that the balls, M M, which are insignificant in comparison with the size of the earth, can only exert an extremely feeble attraction. So small is this that it can only be detected when the beam is entirely inclosed in a case to protect it from draughts; when, further, the whole apparatus is placed in a room into which no one must enter, because the heat of the body would warm the case unevenly, and so set up air currents which would have far more influence than the whole attraction to be measured; and when, finally, the period of oscillation is made very great, as, for instance, five to fifteen minutes. In order to realize how small must be the force that will only just produce an observable displacement of the balls, mm, it is sufficient to remember that the force which brings them back to their position of rest is the same as the corresponding force in the case of a pendulum which swings at the same rate. Now a pendulum that would swing backwards and forwards in five minutes would have to be about 20,000 metres long, so that in this case a deflection of one millimetre would be produced by a force equal to 1/20,000,000 of the weight of the bob. In the case of a pendulum swinging backwards and forwards once in fifteen minutes the corresponding force would be nine times as small, or 1/180,000,000 of the weight. In spite of the very small value of the constant of gravitation, Cavendish was able, by making the apparatus on this enormous scale, to obtain a couple which would produce a definite deflection against the torsion of his suspending wire. These measures were repeated by Reich (Comptes rendus, 1837, p. 697), and then by Baily (Phil. Mag., 1842, vol. xxi. p. 111), who did not in any important particular improve upon the apparatus of Cavendish, except in the use of a mirror for observing the movements of the beam. Cornu and Baille (Comptes rendus, vol. lxxvi. p. 954, vol. lxxxvi. pp. 571, 699, 1001) have modified the apparatus with satisfactory results. In the first place they have reduced the dimensions of all the parts to about onequarter of the original amount. Their beam, an aluminium tube, is only metre long, and it carries at its ends masses of pound each, instead of about 2 pounds, as used by Cavendish. This reduction of the dimensions to about one-quarter of those used previously is considered by them to be one of the advantages of their apparatus, because, as they say, in apparatus geometrically similar, if the period of oscillation is unchanged, the sensibility is independent of the mass of the suspended balls, and is inversely as the linear dimensions. I do not quite follow this, because, as I shall show, if all the dimensions are increased or diminished together, the sensibility will be unchanged. If only the length of the beam is altered and the positions of the large attracting masses, so that they remain opposite to, and the same distance from, the ends of the beam, then the sensibility is inversely as the length. This mistake-for mistake it surely is-is repeated in Jamin's "Cours de Physique," tome iv. ed. iv. p. 18, where, moreover, it is emphasized by being printed in italics. The other improvements introduced by Cornu and |