the left-hand page, and so above the plates, which are immediately under the pupil's eyes. The printing and the plates (the only figure that does not please us is the oval on Plate II.) leave nothing to be desired.

66

We proceed to point out a few matters which we think admit of improvement. Plate II. in the definition of a circle invarying is used; why not constant ?" The construction of Fig. 6 (Plate IV.) is hardly satisfactory to our view, though it is one very frequently given; the tangent to the two arcs is not obtained by a legitimate method. We cannot make out the definition of an harmonic mean given on Plate VII., but the means are correctly constructed. In Fig. 31 (text), for GH: HA, read vice versa. We may remark that it is a curious fact that the approximative construction given in Fig. 87 is true in the cases of regular figures of three, four, and six sides. In Fig. 99 (text) read "through F and E." In Fig. 112 (text) arcs "cutting in C," not G. Constructions to Figs. 123, 125 give particular ellipses; so in the case of the parabolas in Figs. 138, 139, we note that certain figures are stated to be co-centric and certain curves have assym ptotes. In Fig. 271 (text) read to cut in "I and H." We object, on pure geometric grounds, to the constructions in Figs. 278, &c., where a line is found equal to the semicircumference of a circle, &c.; also the inscribed circle of a square and the inscribed triangle are stated as being in the ratio, triangle: circle: square, as 2: 3: 4. In Fig. 279 (text) the two last A's should be D. The construction to Fig. 297 (to draw a line to bisect any triangle from a given point within it) is new to us, and on a cursory examination of it we have not satisfied ourselves of its correctness. In Fig. 314, for X Y, read ZY. In Fig. 316, "the square on," or some such words have been omitted. In Fig. 323 the limitations have not been laid down. In Fig. 329, "join point x," &c.; in 331, for "rectangle " read "parallelogram." These trivial oversights will serve to show how correctly the text has been printed.

OUR BOOK SHELF

66

Observaciones Magneticas y Meteorologicas del Colegio de Belen de la Compañia de Jesus en la Habana, 1873 y 1874. (Habana, 1874 and 1875.)

THE observations made at the College of the Society of Jesus, Havana, are peculiarly valuable for the fulness and care with which they are made, and for the completeness with which the observations themselves and the monthly means and extremes are given in each monthly table and its accompanying diagram. The diagrams, which have been published in their present improved form since June 1873, and which exhibit on one sheet the two-hourly observations as made daily from 4 A.M. to 10 P.M. of all the meteorological and magnetical elements, will very much facilitate the study of those inquiries which deal with the inter-relations of these elements. To these observations are added the daily amounts of the rainfall and evaporation-the latter being of great interest as contributing to our knowledge of the evaporation in intertropical regions, of which so little is known. Whilst only the daily amounts of the rainfall is given, each hour during which rain falls is noted, together with the hour of occurrence of thunder and other irregularly recurring phenomena. As regards the diurnal variations of the wind it changes from about S.E. in the early morning, through E. and N.E. to N.N.E. its most northerly point, which is usually reached about 2 P.M., and thence in the

reverse direction through N.E. and E. to E.S.E., which is reached about 10 P.M. The diurnal velocity is at the minimum at 4 A.M., rises to the maximum at 2 P.M., and thence falls steadily to the minimum. The N. and N.E. winds are decidedly the strongest, and the S.E. the weakest, the ratio being as two to one; in other words, the sea-breeze blows with double the velocity of the landbreeze at this station.

LETTERS TO THE EDITOR

[The Editor does not hold himself responsible for opinions extressed by his correspondents. Neither can he undertake to return, or to correspond with the writers of, rejected manuscripts. No notice is taken of anonymous communications.]

Blowpipe Analysis

66

MR. HUMPIDGE (vol. xiii. p. 208), on the entirely gratuitous assumption that I use commercial reagents"-whatever that term may mean-says that there is probably iron in my soda.

To this I only reply that I will undertake to show pyrologically the presence of 001 per cent. of iron oxide in a fragment of a salt the size of a pin's head; and that, when Mr. Humpidge can do as much without using the dangerous test potassium ferrocyanide (which itself contains iron), I will admit his right to assume that he knows his tools better than other workmen.

No one has ever doubted the proportional relativity in precipi. tating power between a drop and a gallon of water, but if Mr. Humpidge will only do me the justice not to mutilate my statements in the reproduction, he will repeat that a precipitate could not be shown in a drop of water "on a fused mass upon an aluminium plate." W. A. Ross

Shepherd's Bush, W., Jan. 14

The D-line Spectrum

WILL Prof. Stokes give us the reason of his now holding that his first-to all appearance, extremely rational-conclusion, that, in consequence of "the powerful affinities of sodium, it could not cxist in a free state in the flame of a spirit-lamp," is "erroneous"? Shepherd's Bush, W., Jan. 8 W. A. Ross

The Difference of Thermal Energy transmitted to the Earth by Radiation from different parts of the Solar Surface.

THE tenor of certain letters received from scientific persons on the above subject induces me to lay the following statement before the readers of NATURE:

1. Previous to undertaking a systematic investigation of the mechanical properties of solar heat, I examined thoroughly the merits of Laplace's famous demonstration relating to the absorptive power of the sun's atmosphere, proving that only onetwelfth of the energy developed by the sun is transmitted to the earth. The demonstration being based on the assumption that the sun's rays emit energy of equal intensity in all directions, my initiary step was that of testing practically the truth of that proposition. It has been asserted that Laplace did not propound feel called upon, before proving its unsoundness, to quote the the singular doctrine involved in such a proposition, I therefore words employed by the celebrated mathematician. (See "Méchaniquc Céleste," tome iv. page 284.) Having called attention to the fact that any portion of the solar disc as it approaches the limb ought to appear more brilliant because it is viewed under a less angle, Laplace adds :-" Car il est naturel de penser que chaque point de la surface du soleil renvoie une lumière égale dans tous les sens." Let a bed, in the annexed diagram, Fig. 1, represent part of the border of the sun, and b a, cd, small equal arcs; aa, bb, ce, dd', being parallel rays projected towards the earth. Laplace's theory asserts that owing to the concentration of the rays the radiation emanating from the portion de portion of cd to fc. The proposition is thus stated in "Métransmits greater intensity towards the earth than ba, in the prochanique Céleste": "Call the arc of a great circle of the sun's surface, included between the luminous point and the centre of the sun's disc, the sun's radius being taken for unity; a very small portion a of the surface being removed to the distance

from the centre of the disc, will appear to be reduced to the space a cos; the intensity of its light must therefore be increased in the ratio of unity to cos e."

2. In order to disprove the correctness of the stated demonstration, I have measured the relative thermal energy of rays projected in different directions from an incandescent metallic disc, by the following method:-Fig. 2 represents section of a conical vessel covered by a movable semi-spherical top, the vessel being surrounded by a jacket through which water may be circulated. A revolving circular disc, a a, composed of cast iron, the back being semi-spherical and protected by fire-clay, is suspended across the top of the conical vessel supported by horizontal journals attached at opposite sides. The angular position of the disc is regulated by a radial handle, b, connected to one of the journals; the exact inclination to the vertical line being ascertained by means of a graduated quadrant, d. An instrument, c, capable of indicating the intensity of the radiant heat transmitted by the incandescent disc, is applied at the bottom of the conical vessel. The mode of conducting the experiment is extremely simple. The movable cover and its lining of fire-clay having been removed, the cast-iron disc is heated in an air-furnace to a temperature of 1,800 F. It is then removed by appropriate tongs, and suspended over the conical vessel, the lining and cover being quickly replaced. The temperature, shown by the instrument at the bottom of the conical vessel, resulting from the action of the radiant heat of the disc, is then recorded for every tenth degree of inclination. The inves tigation, it may be briefly stated, shows that the temperatures imparted by radiation to the recording instrument is exactly as the sines of the angles of inclination of the disc. Hence, at an inclination of 10° to the vertical line, the temperature imparted to the thermometer is scarcely one-sixth of that imparted when the disc

temperature of 27° 49 F. to the thermometer g, while the radiation from the zone A, Fig. 6, imparted only 619 F. to the thermometer k. Let us bear in mind that the radiating surface / m of the zone A is equal to the radiating surface p q of the zone c. The stated great difference of temperature produced by the radiation from zones of equal area furnishes additional proof that Laplace based his remarkable analysis on false premises. "The sun's disc ought to appear more brilliant towards the border, because viewed under a less angle," we are told by the great analyst. The instituted practical tests, however, prove positively that the energy of the rays projected from the border of an incandescent sphere is greatly diminished because viewed under a less angle from the point occupied by the recording

thermometer.

faces the thermometer at right angles; yet in both cases an equal amount of surface of an equal degree of incandescence is radiating towards the instrument! Laplace and his followers have evidently overlooked this important and somewhat anomalous fact, proving that radiation emanating from heated bodies is incapable of exerting full enegy in more than one direction. Our practical experiments with the revolving incandescent disc have thus fully demonstrated the truth of the proposition intended to be established, namely, that the rays emanating from incandescent planes do not transmit heat of equal energy in all directions, the energy transmitted being as stated, proportionate to the sines of their angle of inclination to the radiating surface.

3. The next step in the investigation of solar heat, before adverted to, was that of measuring the radiant energy transmitted in a given direction by an incandescent solid metallic sphere. For this purpose I employed a double conical vessel similar to the one represented in Fig. 2, the incandescent sphere being suspended over the conical vessel in the same manner as the revolving disc. The nature of the arrangement will be readily understood by inspecting the annexed diagram, which represents four spheres, Figs. 3, 4, 5, and 6, each sphere being divided into four zones, A, B, C, and D, occupying unequal arcs, but containing equal convex areas. Semi-spherical screens composed of non-conducting substances were applied below each sphere, provided with annular openings, arranged as shown in the diagram. Through these annular openings the radiant heat from the incandescent zones, D, C, B, and A, was transmitted to the thermometers, f, g, h, and k, respectively. Père Secchi, and other followers of Laplace, will be surprised to learn that when the suspended sphere was maintained at a temperature of 1,800° F., the radiation from the zone C, Fig. 4, imparted a

4. The result of our experiment with the revolving/incandescent disc shows that if the small arc ba, in Fig. 1, be reduced until the field represented by ba' becomes equal to the field represented by dd', the radiant energy transmitted through each of those fields will be alike; the reason being that the number of rays of diminished intensity passing through c' d' will be as passing through a', as cd is greater than the reduced & a = fc. much greater than the number of rays of maximum intensity

It should be observed that c d is so small that we may without appreciable error regard it as a straight base, and fe as the sine of the angle cdf. It follows from this demonstration that if the solar atmosphere exerted no retarding influence, the radiant heat transmitted towards the earth would be alike for equal areas of the solar disc-more correctly, for areas subtending equal angles, since the receding part of the solar surface is at a greater distance from the earth than the central part.

Encouraged by the practical result of the instituted investiga

tion, I devised the method described in NATURE (vol. xii. p. 517), showing that the polar and equatorial regions of the solar disc transmit radiant heat of equal intensity to the earth, and that the sun emits heat of equal energy in all directions. Adopting Secchi's doctrine relating to the retardation suffered by calorific rays in passing through atmospheres, viz., that the diminution of energy is as the depth penetrated by the rays, it may also be shown by an easy, calculation based on the result of our investigations, that the absorption by the solar atmosphere cannot

exceed one-seventh of the radiant energy emanating from the photosphere. 5. Concerning the plan resorted to by the Director of the Roman Observatory, and others, of investigating the sun's image instead of adopting the method of direct observations, I will merely observe that the information contained in the several works of the Roman astronomer furnishes the best possible guide in judging of the efficacy of image investigation. Let us select his account of the investigations conducted between the

19th and 23rd of March, 1852. Having pointed out that in these experiments it was impossible to approach within a minute of the edge of the sun, and that during a later observationdate not mentioned-he had approached within a minute, the investigator observes: "But at this extreme limit, even making use of the most accurate means of observation, we find difficulties which it is impossible to overcome completely." In addition to this emphatic expression regarding the difficulties encountered, the author adds: "Moreover, it is impossible to study the edge alone, for the unavoidable motions of the image do not admit of its being retained at exactly the same point of the pile; we have therefore been unable to push the exactness as far as we hoped; and we have discontinued the pursuit of these researches, although the results obtained are quite interesting." (See revised edition of

OUR ASTRONOMICAL COLUMN STAR WITH SUSPECTED LARGE PROPER MOTION.It would appear by a communication from Prof. Winnecke, Director of the Imperial Observatory at Strasburg, that the large proper motion exhibited by a comparison of Argelander's positions of the ninth magnitude star, No. 11237-8 of Oeltzen's catalogue (southern zones) with Taylor's observations at Madras in 1838 or 1839, to which reference was lately made in this column, does not really exist, there being evidently an error in Taylor's mean place for 1840 given at p. clxiii. of vol. v. of the Madras Observations. Prof. Winnecke finds that the differences of right ascension and declination between this star and Oeltzen 11226, are sensibly the same as at the time of Argelander's observations (1851), and the latter star is known to have but very small, if any, proper

"Le Soleil," vol. i. p. 205.) It is needless to institute a comparison between a system of which its founder speaks so despondingly, and one which enables us to push our investigations to the extreme limit of the solar disc, admitting of entire zones being viewed at once, instead of only small isolated spots.

J. ERICSSON

The Glow-worm in Scotland THE Glow-worm is not uncommon on the Island of Cumbrae, Buteshire. I have seen it there occasionally for the last thirty years (see vol. xiii. pp. 188, 208). DAVID ROBERTSON

Millport, Island of Cumbrae, Jan. 18

motion. Taylor's star must therefore be struck off the list of cases of great proper motion lately given.

ATLAS 27f PLEIADUM.-A very interesting observation was made at Strasburg on the occasion of the occultation of this star-a Struve's difficillima-on the 7th of the present month. As we recently stated, this star does not appear to have been seen double since the last Dorpat observation in 1830. On the 7th inst., however, Herr Hartwig observing at Strasburg with an excellent Fraunhofer, of 42 lines aperture, power 159, remarked that the star did not disappear instantaneously; after the principal mass of light had vanished there remained a luminous point for about six-tenths of a second, a circumstance which favours the duplicity of the object, notwithstanding the failure of recent efforts to divide it. It brings to our recollection Burg's observation of the

occultation of Antares 1819, April 13, when at emersion the star appeared to suddenly increase from one of the sixth or seventh magnitude to one of the first, a phenomenon no doubt attributable to the existence of the small companion on the parallel, preceding the principal star (NATURE, vol. xii. p. 308).—The next occultation of Atlas-Pleiadum, on February 3, will not be visible in this country, but may be well observed in the United States. The American Ephemeris gives the time of immersion for Washington; at the Observatory of Hamilton College, Clinton, N.Y., so actively conducted by Prof. Peters, the immersion takes place at 11h. 13m., and the emersion at 12h. 4m., Clinton M.T.

VARIABLE STARS.-In No. 2071 Dr. Julius Schmidt, of the Observatory, Athens, continues his elaborate researches on the three short-period variables U, W, and X Sagittarii, the periods of which are now given thus:

U Sagittarii ...
W

Sagittarii

X = 3 Flam.

S.

d. h. m. 6 17 53 14 7 14 15 34'I 7 0 17 42'5 So assiduously have these stars been watched by their discoverer, Dr. Schmidt, in the fine skies of his locality (little success could be expected to attend their observation in England), that he believes he has detected perturbations of the light curve or period in each instance, though not quite ten years' observations are yet upon record. The following are Greenwich times of geocentric minima of Algol, according to Prof. Schönfeld's ele

ments:

the apartments the Society occupied in Somerset House, it was at once seen that the most formidable work the change involved would be the removal of the collections of minerals and fossils. The transference of the library, though an extensive one, would be a comparatively easy matter, but there is always the danger in the mere handling of fossils that they may be damaged. Besides this, the collection had gradually grown to such a size that it was evident the cost of the removal would be considerable. So far as the preparation of the rooms at Burlington House was concerned, the Government showed every desire to conform as far as possible to the wishes of the Council.

Some of the Fellows'counselled that the whole collection should be offered to the British Museum or to the School of Mines Museum in Jermyn Street, on the ground that though in the early days of the Society it was of high value when it was the only museum that existed, it was now so far surpassed in magnitude by the national collections that it was practically of small value. Fortunately wiser counsels prevailed. There were in the museum, it was urged, many typical collections formed by the early leaders of geological science, which were bequeathed in illustration of papers they had read and work they had done. These collections, obtained by their own personal labour in the field, arranged and named in their own handwriting, were of historical value and had a European reputation, and ought to be religiously preserved by the Society. It was true that the integrity of some of the collections had been destroyed in the endeavour at one time to make one general collection illustrating the whole of England, and arranged in stratigraphical order; but in most cases the original labels and references to catalogues were preserved, and it was hoped it might be possible in the new buildings to regroup the specimens much as they were at first. It was therefore determined that the museum

h. m.

1876. Jan. 29 13 46 Feb. 17 13 2 March 7

12 19

h. m. 1876. April 14 10 54 May 3 10 12 22 9 31

26 11 36 RECENTLY-DISCOVERED MINOR PLANETS. No. 152, discovered at Paris by M. Paul Henry on Nov. 2, has been named Atala, and for No. 157, the small planet, detected by M. Borrelly at Marseilles on Dec. 1, the name of Dejanira is proposed; elements of this planet have been calculated by M. Stephan. The following are first approximations to the positions of the ascending node, inclination, and periods of the newer minors, with dates of discovery :

tion, but mainly as a repository of specimens referred to in papers, and that before the removal commenced it should be carefully weeded, so that in all cases where, through the accidental removal of a label or other causes, the history of any specimen had been lost, it should be discarded, but not until every effort had been made to try to ascertain any possible clue. This work has been carried out by Prof. Rupert Jones, aided by Mr. Woodward, the assistant curator. The accumulation of specimens had caused much crowding in the museum, and in such a case a certain amount of damage and loss of labels was almost inevitable. As a consequence of this weeding, many specimens have been omitted in the new arrangement, and the result has been to leave greater space for those that have a real historic value.

Like many other institutions of gradual growth, the history of this museum has never been written, and very few people, few even of the Fellows of the Society, know what it contains, for there never has been a printed catalogue. As the collections are the private property of the Society and are not open to the public, this perhaps has not been thought requisite.

Among the principal collections preserved which have now historic value, first in point of general interest should perhaps be mentioned the extensive series of fossils presented by Sir Roderick Murchison, from which were drawn the figures in his world-renowned "Siluria." The fossils figured in the papers by Murchison and Sedgwick in describing the structure of Wales and the Lake district are also there, so are the fossils that illustrated Murchison's description of Brora. The fossils connected with Webster's well-known paper of 1814, the first paper on the Tertiaries of Hampshire; most of those illustrating Fitton's celebrated paper on the "Strata below the Chalk" (1827); those belonging to Buckland and Conybeare's "On the South-west of England" comprehensive paper (1824) are all there. Large additions to the general col

lection were also made by Dr. Mantell, Dr. Macculloch, and Mr. Leonard Horner.

It will be recollected that the Society was originated in 1807, at a time when mineralogy was a fashionable study, or at least when collections of minerals formed part of the "furniture" of the apartments of the Queen and many of the nobility. Collections of shells and of fossils were also fashionable, but they were valued only for their beauty or their rarity, and not for any knowledge of nature they afforded. For some time the young society seems to have followed fashion. Indeed, the value of fossil organic remains as giving a clue to the consecutive sequence and relative order of strata was then but just beginning to be understood. It was not till the end of 1799 that the first MS. table of the sequence from the Carboniferous beds upwards was constructed, and no map of the strata of England was published till 1815. The earliest MS. catalogue of specimens belonging to the Society, begun in 1808 or 1809, is labelled "General Catalogue of Minerals," and some of the early entries of organic fossils refer rather to the rock in which the fossil is imbedded; the presence of the fossil being but casually noticed, such as "limestone containing shells." These early collections of fossils illustrating the labours of the first geologists in using organic remains to trace the chronological sequence of beds, and to compile some chapters of the earth's history, have a profound interest, laying as they did the foundations of a science which has placed at rest many wild theories of the origin of the earth, and has, too, proved to be of such practical value. The first donation recorded is Feb. 5th, 1808, of specimens from St. Anthon's Colliery, Newcastleupon-Tyne, by the Right Hon. Sir J. Banks. It would Occupy too much space to mention all the collections that the Society has preserved, but among the donors are the well-known names of Sir Henry de la Beche, Sir Charles Lyell, Greenough, Warburton, and Sir Woodbine Parish. McEnery's collection that first brought Kent's Cavern into notice is there, and so is a splendid series of Daniel Sharpe's "Brachiopoda." The old red sandstone fishes presented by Lady Gordon Cumming are remarkable for their beauty as well as for the extent of the collection. Many distinguished living geologists have private collections of their own; for example, the Earl of Enniskillen, Sir Philip Egerton, Prof. Prestwich, Mr. Searles Wood, Dr. Bowerbank, &c., which fully explains why their contributions are not so numerous as might be expected from the valuable work they have done. Prof. Phillips, though so energetic a worker, is not largely represented in the museum, for firstly York, and afterwards Oxford, had stronger claims on him. The same remark applies somewhat to the claims of the Woodwardian Museum on Prof. Sedgwick. As illustrating the geology of England generally, the Jermyn Street Museum and the British Museum are more useful, but as a record of early geological work the museum of the Society is unique.

The rearrangement of the foreign collections has not yet been completed, though it is in progress. Suites of specimens are to be seen there from all parts of the known world from which it has been possible for travellers to send them. These foreign collections are, to some extent, the result of contributions by officers in Her Majesty's services. Central Africa is not represented, but there are several collections from both coasts. For the future it is intended to add to the British collection only those specimens that are sent in illustration of papers read to the Society, but foreign specimens will be received as before.

Among the treasures of the museum, besides the rocks and fossils, there are the original drawings of Agassiz's "Poissons Fossiles," presented by the Earl of Enniskillen, the first manuscript geological map of England (1799), and the first table of strata, by W. Smith (1799).

The previous changes in the locality of the museum have

been as follows:-In No. 4, Garden Court, Temple, the first fixed habitation of the Society (June 1809), the collection was commenced. In June 1810 it was removed to 3, Lincoln's Inn Fields; in June 1816 to 20, Bedford Street; in the autumn of 1828, to Somerset House; at Somerset House it has remained till this last move to Burlington House.

FOR

CONDENSED AIR TRAMWAYS

OR some weeks the North Paris Tramways Company has been trying on the line from Courbevoie to the Arc de l'Etoile a new system of locomotion, in which the motive power is compressed air. Some details of M. Mékarski's (the inventor) system are given in the Revue Scientifique. It is capable of considerable developments and of varied applications, since it has solved in a very satisfactory manner the double problem of the industrial production of air condensed to very high pressures, and of the storage of the air in reservoirs intended to discharge into a cylinder placed in any apparatus whatever, at any distance from the condensing pump. The "Voiture Automatique" of M. Mékarski is characterised by the absence of an imperial and by a platform in front and another behind. This car carries the reservoirs of condensed air, the apparatus for distribution, and the cylinders. M. Mékarski places under the truck of the car the sheet-iron cylinders, which contain the condensed air; on the front platform is placed the distributing apparatus which the engine-man works; the two cylinders are placed, as in certain locomotives, outside the framework, horizontally, and act directly, by means of a crank, on the front wheels of the car. No doubt this arrangement might be advantageously modified; but the important point is the possibility of manufacturing compressed air in sufficient quantities to be of use as a motive power.

The condensing apparatus used by M. Mékarski consists of two pump-barrels of cast-iron, placed vertically, communicating respectively with two horizontal pumpbarrels, in which move two pistons worked by a steamengine. This is, in reality, a double condensing pump, the former bringing the air to the pressure of from ten to twelve atmospheres, and the second raising the pressure to twenty-five atmospheres. The pistons act upon a mass of water which compresses the air directly and absorbs by degrees the heat disengaged by compression. By an ingenious contrivance the supply of water is continually renewed, and the temperature thus kept down. But this arrangement does not absorb a sufficient amount of the heat disengaged, a difficulty which M. Mékarski has met as follows. The external air drawn into the pump raises a valve constantly covered by a layer of water of several centimetres; besides, a large cast-iron tube, constantly traversed by the air already condensed and the excess of water, communicates with the two vertical pump-barrels ; finally, the second pump is fitted with a tap by which the heated water escapes.

In M. Mékarski's automatic car the compressed air is stored, under the truck, in sheet-iron reservoirs or cylinders. The total capacity is about 2,000 litres; 1,500 litres serve as an ordinary supply; 300 litres constituting a reserve; the remaining 200 litres are intended to serve as a brake. The air is compressed in the cylinders to the pressure of twenty-five atmospheres. On the line from Courbevoie to the Arc de Triomphe, 7,500 metres going and returning, the resistance is unusually great. In one experiment the ordinary feeding cylinders contained 1,500 litres of twenty-five atmospheres at departure, and the pressure, on arrival, was not more than four and a quarter atmospheres. The expenditure had thus been about 1,250 litres at twenty-five atmospheres for a run of 7,500 metres, or 166 litres per kilometre.

But unless it is possible to heat the air gradually