complex these structures, the more convincing the evidence they furnish, since it then becomes so much the less probable that the same anatomical device should have been produced twice than that it should have been acquired once only. In the Artiodactyla, for example, "a definite case of an anatomical peculiarity, so well marked and complex as to be a safe guide to affinity," is seen in the elongation and spiral coiling of the proximal portion of the colon. The Perissodactyla and rodents supply other examples of evolution along a definite radius from the ancestral centre. From his investigations the author deduces inferences of importance for the general theory of evolution, especially as regards the limitation of the possible range of variation of organs in any set of animals which have once come to Occupy a particular radius. Further changes and elaborations are then restricted by the past history, that is to say, by the limited material which it has furnished for further specialisation. In this way a simple explanation is given for the definite grooves, recognised by many writers, along which the specialisation of organisms must necessarily move, without having recourse to the assumption of any mysterious directive forces. E. A. M.

SCIENTIFIC REPORTS OF THE LOCAL

GOVERNMENT BOARD.1

THE first half of the volume under notice is devoted to

the medical officer's report, statistical data, and details of various inspections and inquiries by the Board's inspectors. The second half contains reports of the auxiliary scientific investigations carried out for the Board. The first of these is a memorandum by Dr. Theodore Thomson on rats and ship-borne plague. The conclusion arrived at is that "the part played by the rat in transmission of plague to man, although real, falls far short of the importance which has generally been attributed to it." This may be true, but in view of the predominant part played by the rat in the dissemination of plague in the various Sydney epidemics, it is to be hoped that the campaign against this rodent will in no way be relaxed.

Bearing on the same subject, Drs. Haldane and Wade report on methods of rat destruction and disinfection on ship-board. In this especial attention is directed to the Clayton process, in which sulphur is burned at a high temperature, and air charged with the products of its combustion is pumped into the ship's hold. The gas is rapidly fatal to rats and other vermin, and is germicidal to non-sporing microbes, but it does not penetrate a loaded hold well, and has a deleterious action on certain articles. On the whole, however, it seems to be the best method to employ for rat destruction. Dr. Klein details further experiments on the two types, virulent human and less virulent rat, of the plague bacillus differentiated by him and described in a previous report. Dr. Klein also records some interesting observations on the influence of symbiosis on the virulence of microbes.

the

An important paper on the differentiation of various streptococci and staphylococci is contributed by Dr. M. H. Gordon. Hitherto the differences exhibited by the members of these classes of micro-organisms, particularly streptococci, have been slight and indefinite, but by making use of culture media containing various mono-, di-, tri-, and poly-saccharides and glucosides, important differential characters are obtainable. Dr. Sidney Martin has continued his studies on the toxic action of microbes, dealing in the present volume with that of the Proteus vulgaris. The results, however, in this case are somewhat indefinite, the toxic reaction being mainly evinced by the development of agglutinin in the blood. Dr. Houston gives a detailed report of the bacteriological examination of normal human dejecta, and of the intestinal contents of sea-fowl and of fish. All gulls contained typical B. coli in their excrement in enormous numbers, but guillemots did not contain B. coli of any sort. As regards fish, those obtained from a source seemingly above all suspicion of objectionable contamination, may contain sometimes apparently typical B. coli in their interior; in the great majority of 1 "Thirty-third Annual Report of the Local Government Board, 1903-4.' Supplement containing the Report of the Medical Officer for 1903-4

enormous

THE record of an amount of work on the absorption spectra of organic compounds and emission spectra of various metals and gases in the infra-red region is given in the volume under notice. The investigations were commenced whilst the author was a graduate student at Cornell University, and completed under the auspices of the Carnegie Institution of Washington.

Even to summarise the mass of valuable information contained would exceed the limits of our present space, but it may be said at once that, to workers along similar and related lines, these results, and the descriptions of the apparatus and methods whereby they were obtained, are indispensable.

Part i. occupies nearly seven-eighths of the whole volume, and deals with the absorption spectra of 131 organic compounds up to 15 μ. As is pointed out in the very complete historical review, all previous workers in this subject have abandoned the investigation at 7μ for the alcohols and 10 μ for some few other compounds.

The description of the apparatus and methods is exhaustive and invaluable. From 08μ to 2.5 μ a quartz prism was employed, beyond that, and up to 15 μ, one made of rock-salt. The source of the radiations was a Nernst lamp "heater," which gives a spectrum of which the energy curve is smooth and continuous. A reflecting spectrometer of 35 cm. focal length was employed for the explorations of the spectrum up to 15 μ, and a considerable portion of the work up to 7.5 μ was repeated with a spectrometer of 1 m. focal length.

The distribution of the energy in each spectrum was determined by means of a radiometer similar to that devised by Nicholls, but with some modifications.

The principal reasons for this investigation were the determination of the influence of molecular weight upon absorption spectra, and also the effect of molecular structure. The results show that in different compounds each of these causes in turn acts separately, whilst in other compounas the absorption is produced by the combined effect.

In recording the quartz-prism results the author deals separately with each absorption band in the nineteen compounds investigated, whilst in the other results the compounds are treated separately, notes being made of the chemical structure and properties of each substance where

necessary.

Numerous tables set out the numerical results in various forms, whilst 140 full-page transmission curves show them graphically. In addition to these the author has written seven brief appendices dealing with side-issues in connection with the apparatus and the investigation and its results.

In part ii. Mr. Coblentz deals with the infra-red emission spectra of various metals, metallic chlorides (alkalies), and gases. The metals were employed as the poles for the arc producing the radiations, whilst the chlorides were volatilised on carbon arcs. The apparatus was very similar to that described in part i., except for a few modifications rendered necessary by the greater intensity and unsteadiness of the radiations.

With the metals, a black-body spectrum due to the oxides, and sufficiently strong to obliterate any emission lines which might be present, was found, and in the alkali chloride spectra no lines were discovered beyond 2 μ. Of the gases investigated-in vacuum tubes-N was found to be the only one having strong emission lines in its infrared spectrum. CO,, CO, and the vapour of C,H,HO were

1 "Investigations of Infra Red Spectra." By William W. Coblentz. Pp. vi+331. (Washington, D.C.: The Carnegie Institution, 1905.)

found to exhibit a very strong emission band at 475 μ. The emission spectrum of C,H,OH shows that a vapour in a vacuum tube can emit a continuous spectrum.

Angstrom's conclusions-deduced from the fact that the total radiation increases, while the luminous radiation decreases, with increase of pressure in the gas-that there are two kinds of radiation present during the electrical discharge are found to be in close agreement with the observed facts. These different discharges were named " regular and irregular" (i.e. luminescence) by the previous observer. An interesting theoretical discussion of the action of pressure in this connection is given in the volume. W. E. R.

66

DISEASES OF VINES.1

TWO parts of the Annales de l'Institut Central Ampelologique Royal Hongrois, devoted to two of the vine diseases, have lately reached us.

In one of these parts (part iii.) an account is given of the little known disease caused by the attacks of Phyllosticta Bizzozeriana. The disease was first noticed in the year 1900, and it has been kept under observation since then, with the result that its spread has been traced in some detail. The symptoms are somewhat similar to those of the dreaded "black-rot," but it does not appear as if it will prove so dangerous a parasite. In addition to a brief life-history of the fungus, illustrated by an excellent plate, a useful compendium of the species of Phyllosticta occurring or the vine is given.

66

Part iv. contains an unusually complete account of the grey-rot" caused by Botrytis cinerea. This is one of the parasites of the vine which the cultivator most dreads. All aerial parts of the host-plant are attacked indiscriminately, and quickly become covered with a greyish or brown mould, which produces enormous quantities of ashy-grey spores. This stage is succeeded by the formation of small black sclerotia in the diseased tissues of the stems, leaves, and fruits. Naturally the fungus has been investigated time after time, but the researches of Istvanffi, published in this volume, have added a number of fresh facts to our knowledge of its life-history. In the first place, a series of laboratory investigations was made with the object of determining the conditions under which the fungus brought about the infection of the host-plant. The optimum temperature for the germination of the spores proved to be 25° C., whilst the spores were killed by exposure to a temperature of 38° C. to 41° C. The effects of drying the spores were then investigated. One day's drying over sulphuric acid at laboratory temperatures, either in light or darkness, was sufficient to kill 75 per cent. of the spores, and desiccation for thirty-six days was fatal to all of them. Spores previously germinated and exposed to this treatment suffered still more severely. The results of freezing were again seriously to diminish their germinating capacity.

The action of a number of the commoner fungicides on spores was then examined, with interesting results. Thus a I per cent. solution of Bordeaux mixture only prevented the germination of some 60 per cent., and a 10 per cent. solution about 10 per cent. Spores which were allowed to dry after soaking in Bordeaux mixture all failed to germinate. Others sown in drops on the foliage of the host-plant not only germinated, but infected the tissues below them. On examining the action of the constituents of this mixture, lime water proved to be singularly efficacious in preventing germination.

In the majority of these experiments the spores of Monilia and of Coniothyrium were exposed to the same conditions, with results, on the whole, similar to those already quoted. At the same time, the life-history of the fungus was traced in detail. Istvanffi succeeded in germinating the sclerotia, and has settled the point once for all that they do give rise to the apothecia of Sclerotinia fuckeliana. So many observers have failed to obtain this ascigerous stage that it is well to have this definite statement. The sclerotia retain their germinating capacity for at least twenty-one months. Another interesting point "Annales de l'Institut Central Ampélologique Royal Hongrois," tome iii., livre 3 and 4, 1905. (Budapest, 1905.)

brought out in the course of this research is that the well known adpressoria of the fungus are the early stages in the development of the sclerotia. In addition to the microconidia observed by Brefeld and others, Istvanffi records the production of an oidial stage.

For further details, and for methods to be adopted for checking the spread of this pest, the original must be consulted. It is full of points of interest to the student of plant pathology, and makes one regret more and more that this country possesses no institute similar to the Central Ampélologique Royal Hongrois, where the pressing problems of plant disease can be adequately examined. Here we have to trust to the private individual for what investigations are made, and he all too rarely has opportunities to make them on the comprehensive scale possible at such an institution.

EARTHQUAKE ORIGINS.

AMONG the most interesting and important of the new ideas, which have been introduced into seismology, in late years, must be classed Major E. G. Harboe's notion of the nature of earthquake origins. Originally treated as a point, the focus of an earthquake has long been recognised as an area, but we are still in the habit of regarding it as restricted in size and small in comparison with the dimensions of the area over which the earthquake is felt. On this hypothesis the decrease in violence is correlated with increase in distance, and due to a gradual diminution of intensity as the disturbance travels from its origin. according to Major Harloe's conception, the focus of an earthquake is no longer restricted in size, but ramifies, with a varying degree of initial violence, over nearly the whole of the seismic area.

On the generally accepted hypothesis the coseismal lines should more or less correspond with the isoseismal, a decrease in violence being accompanied by an increase c time interval, but such is far from being the case, and w have been in the habit of attributing the irregularities to errors of observation; Major Harboe has now shown that another explanation is possible, and that the irregularities in recorded times almost disappear if his hypothesis of the nature of the origin is adopted. From the discussion of the records of earthquakes he reaches the conclusion tha the true rate of propagation of the sensible shock is about 0.4 kilometre per second, the higher velocities obtained by other investigators being compounded of the rate of propagation of the disturbance along the origin, and that c the wave-motion set up by this disturbance.

This rate of propagation is that of the sensible shock, which can be felt by human beings, and not that of the large waves recorded by seismographs outside the seismic area proper; the latter, the rate of propagation of which is about 3 kilometres per second, are regarded as different in character, and propagated in the consolidated rock at some little depth below the surface, the sensible shock being due to quite superficial waves propagated through the more fissured and less coherent surface rocks.

One of the weightiest of the objections to this hypothesis was the value of 3.28±0.05 km. sec. obtained by Prois. Sekiya and Omori in 1902 from the seismic triangulation started by Prof. Milne in 1884. This is dealt with in vol. viii., part iii., of Gerland's Beitrage zur Geophysic where Major Harboe remarks that, in spite of the leng period over which the observations extended, only four earthquakes seem to have given usable records from al the stations, three earthquakes at three stations, and one at two stations. Taking two of these earthquakes, for which records from a number of meteorological observ atories have been published, he finds that the velocity and direction of propagation, deduced from the triangulation. lead to most discordant results at other stations, irregular ties which disappear if a branch of the origin is supposed to have traversed the field of triangulation and the disturb ance to have spread outwards to the stations.

Whether the hypothesis stands the test of future investigation or not, it seems to explain many previously inexplicable anomalies, apart from those of time. appears to work out satisfactorily in the case of those earthquakes by which Major Harboe has attempted to test

it, but the only true testing must be left to investigators of the future, for it is to be feared that in the past seismologists have been inclined to reject, as bad, all records of time which failed to fit in with their preconceived ideas of the direction of propagation of the shock, though they might have fitted in with a less simple, though possibly truer, conception of the form and extent of the earthquake origin. R. D. OLDHAM.

CURRENTS IN THE STRAITS OF MESSINA.

FOR our knowledge of the physical conditions at the

bottom of the sea we are very largely beholden to the enterprise of submarine cable companies; indeed, it is difficult to imagine a more thoroughly satisfactory method of survey than that employed by them. Duties connected with the maintenance of cables have led to the discovery of details in the configuration of submarine gullies, of freshwater outlets beneath the sea, and of alterations in the bed of the ocean itself, which would otherwise have eluded observation. Prof. Platania, of the Istituto Nautico of Catania, has directed attention to another rather surprising fact, namely, that in the Straits of Messina there are deepwater currents of sufficient velocity to cause the interruption of the cables joining Sicily with the mainland (“I cavi telegrafici e le correnti sottomarine nello stretto di Messina," reprinted from the Atti della R. Accademia Peloritana, vol. xx.). The period under observation covers the last forty years, during which time there have been twenty-six interruptions; neglecting two, nineteen occurred between November and April, and five between May and October. The strong currents cause a continual attrition by sand and pebbles. The rocks on the sea bottom are swept free of mud and sand, and their rough surfaces, thus exposed, have worn out the cables lying upon them. In one case a cable seems to have been corroded by a sulphurous spring. The surface currents attain a speed of five miles an hour. They have always been a danger to navigation, and the wrecks of two large vessels which were lying last summer upon the Sicilian shore show that Scylla and Charybdis have lost none of their power. The existence of correlated strong deep-water currents had been suspected. Biologists have long been attracted to Messina by the plentiful harvest of deep-sea animals which are occasionally brought up to the surface by a vast turmoil of waters, thus affording almost unique opportunities. M. Thoulet and others have repeated the classical experiments of our countryman, Captain Richard Bolland, made in 1675 in the Straits of Gibraltar, and have demonstrated the existence, at twenty fathoms, of an undercurrent flowing in a contrary direction to that on the surface, but these currents have not yet been systematically studied as the importance of the subject demands. The tides, as is frequently the case in narrow straits, as, for instance, inside the Isle of Wight, are doubled.

A PERIODICAL FOR PALEONTOLOGISTS.

as

THIS new venture in scientific literature, which is to

appear quarterly, and leads off with a double number, will be warmly welcomed by all palæontologists, for since the Annales des Sciences Géologiques ceased to exist, there has been no accredited journal for palæontology in France. The "Annales des Sciences Naturelles : Zoologie,' it is true, has on occasion offered the hospitality of its pages, but the whole of its space is not too great for the living subject.

Material enough and to spare lies ready to hand at the Paris Museum in collections from all parts of France and its colonies, while it is further intended to carry on D'Orbigny's incompleted tasks begun in his " Paleontologie Française" and Prodrome de Paléontologie stratigraphique universelle." The publication of illustrations of the yet unfigured types of the latter work, with reprints of the author's diagnoses, accompanied by notes and ex1 "Annales de Pa'éontologie, publiées sous la direction de Marcellin Roule. Tome i. fasc. 1 and 2, January, 1905. Pp. xi+100; 9 plates. (Paris: Masson et Cie.)

planations, an undertaking of great merit, is begun in this first part.

As regards guiding principles, the editor, while not wishing in any way to dictate to his contributors, gently suggests in his introduction that he has preferences. On the one hand, he seeks memoirs on stratigraphical or purely systematic palæontology, in which the principal object will not be the multiplication of genera and species, holding as he does that mieux valent des choses sans noms que des noms sans choses. On the other, he inclines to papers having a philosophic bearing.

With his former predilection all must be in accord, while of the latter, the very first paper, one by the veteran Albert Gaudry, "Fossiles de Patagonie. Les attitudes de quelques Animaux," is an excellent example, where “attitudes is used to express the comparative bearing, gait, and appearance, and not posture alone. The author points out that in Tertiary times in Patagonia Plantigrades and Rectigrades predominated over Digitigrades.

The editor and M. A. Thevenin give the first instalment of a series of memoirs on the palæontology of Madagascar, in which they deal with the molluscan fauna from newly discovered Upper Cretaceous beds on the eastern side of the island. Some of the species enumerated are identical with those found by the Rev. R. Baron in the northern and north-western districts, that were described by Mr. R. B. Newton in the Quart. Journ. Geol. Soc. for 1889 and 1895, a fact to which, however, allusion is not made. This fauna presents considerable analogy with that which lived during the same epoch in India.

The second contribution to the same series, by M. Douville, treats of some nummulitic beds in Madagascar. M. Boule adds a memoir on "Les grands Chats des Cavernes," principally the lion, that takes the form of a popular review of current knowledge on the subject.

The part concludes with the opening portion of the descriptions and figures of D'Orbigny's types already referred to.

Altogether there are 100 pages of text, with nine phototype plates, besides abundant illustrations in the text, all the figures being most excellent, and veritable works of

art.

There is, indeed, but one objection to raise, and that having its distinct pagination in addition to that of the is against the adoption of dual pagination, each paper volume, because the disadvantages of this system for purposes of citation far outweigh any possible benefits.

It is to be hoped that the glossy surfaced paper selected, so suitable for modern text illustrations, though not for type of the face employed, is not of that perishable description which we have been lately warned will deprive future generations of the fruits of our intellectual labours. B. B. W.

UNIVERSITY AND EDUCATIONAL INTELLIGENCE.

report

CAMBRIDGE. The special board for mathematics has put forward new proposals, both with regard to the mathematical tripos and the mechanical sciences tripos, which involve far-reaching changes. The first-named points out that in the opinion of the special board the existing mathematical tripos is unsatisfactory as an examination. The special board proposes to substitute for the present part i. a new part i., which may be taken by a student either at the end of his first or second year. Part i. will not qualify for a degree without further examination. It is hoped that this part will be taken by many who propose to proceed later to study engineering or natural sciences. The board further proposes that for the existing part ii. a new part ii. be established, which must be taken at the end of the third year. The position of senior wrangler is abolished, but the class list of each part will contain three classes, the names in each class being arranged alphabetically. Schedules are published for each of the proposed new parts.

With regard to the report of the mechanical sciences tripos, the special board of mathematics suggests that part . of the tripos should be abolished, and it is proposed to modify part i. by the inclusion of a number of

papers on questions of greater difficulty or of wider range than the average of those now set. The other papers of this part are, however, to be made easier than the present average. The board hopes to include a paper on chemistry in the future. It is also considered to be desirable that the examiners should be empowered to take into consideration the laboratory and drawing-office work done by the student during his course; but perhaps the most important of the recommendations is that every candidate for the mechanical sciences tripos, unless he has obtained honours in one of the honours examinations of the University, must pass a qualifying examination in elementary mathematics and mechanics, which will be held twice a year.

re

The special board for biology and geology has nominated Mr. F. A. Potts, of Trinity Hall, to use the University table at Naples for four months from April 1. Applications for the use of this table and for that at the Marine Biological Association's laboratory at Plymouth should be sent in to the chairman of the special board (Prof. Langley) on or before May 24.

Dr. Haddon is giving a special course of lectures on magic and savage religion on Mondays during this term.

PROF. FRIEDRICH CZAPEK, of the Prague Technical High School, has been appointed professor of botany in Czernowitz University. Prof. Armin Tschermak, of the University of Halle, has been appointed professor of physiology and medical physics in the Veterinary High School, Vienna.

IT is announced by Science that Adelbert College, Western Reserve University, has received 30,000l. from the grandchildren of Mr. Joseph Perkins, formerly a trustee of the college. The money is to be used for a department of sociology and a chemical laboratory.

ON Commemoration Day at Glasgow University on April 18 the honorary degree of Doctor of Laws was conferred upon Mr. James S. Dixon, founder of the lectureship of mining in the University, and Mr. R. E. Froude, superintendent of the Admiralty experimental works at Haslar.

ALTHOUGH We are far behind other nations in governmental recognition of the claims of anthropology, the universities, the older ones leading the way, are following their Continental sisters in making it a subject of systematic study by providing courses of instruction and establishing diplomas and other distinctions. The Oxford committee for anthropology has just issued the regulations for the diploma and the list of lectures for the next two terms. It is pointed out by the committee that not only members of the university, especially those whose work will bring them in contact with native tribes, will benefit from the newly-established course of study, but also those already in contact with native races who feel the need of extending their anthropological knowledge during their "long leave." The schedule of lectures shows that although no provision can yet be made for systematic instruction covering the whole of the very wide field in even a summary manner, students who present themselves are sure of finding helpful and stimulating teaching in all the more important branches of the subject: the chief omission at present is the failure to include social organisation, usually a crux for missionaries and the untrained generally, among subjects on which aid may be sought. The secretary of the committee is Mr. J. L. Myres, Christ Church, from whom all information may be obtained.

SOCIETIES AND ACADEMIES.
LONDON.

Royal Society, Fbruary 1." A Further Communication on the Specificity and Action in vitro of Gastrotoxin." By Dr. Charles Bolton.

An analysis in the test-tube of the gastric cytotoxin obtained by injecting the rabbit with guinea-pig's stomach cells has shown that it is a complex body. After a single injection there is a great increase in the hæmolysin normally occurring in the rabbit's blood, and after further injections an artificial hæmolysin makes its appearance. The artificial hæmolysin is distinguished from the natural hæmolysin, because the former can be complemented by guinea-pig's normal blood serum, whereas the latter cannot.

There is also present in the immune serum a substance which agglutinates the red blood corpuscles. Closely associated with the appearance of this artificial hæmolytic immune body is that of an agglutinin which acts upon the gastric granules, and also that of a precipitin which acts upon the soluble proteids of the gastric cells. By repeating the injections these substances are found to be present in the blood for several months. Whether they are one and the same or distinct bodies has not yet been proved. After several injections, and not less than about five weeks from the first, a further substance appears in the blood, which possesses an action upon the intact gastric cells. In spite of repeated injections this substance disappears from the blood in about four months. It is probably of the same nature as a hæmolysin, but this point requires proof.

or

The hæmolytic factor is only active against blood. The actions of the agglutinin and precipitin are not confined to the constituents of the gastric cells, but extend to other the proteids of body. Whether there are separate agglutinins and precipitins for different proteids, whether the same substances act upon all proteids, has not been determined; at all events, if the same bodies are concerned in all cases, their action upon the proteids of the stomach cells is probably greater than that upon other proteids. Whether the gastrolysin itself is truly specific remains to be proved.

The few experiments that have been undertaken in the case of the human stomach indicate that the human gastric cytotoxin is identical in constitution with that of the lower animals.

February 8." Explosions of Coal-gas and Air." By Prof. Bertram Hopkinson.

The explosion of homogeneous mixtures of coal-gas and air at atmospheric pressure and temperature is investigated by means of platinum resistance thermometers placed at various points in the explosion vessel. The vessel is of dumpy cylindrical form and 6-2 cubic feet capacity, and the mixture is fired by an electric spark at the centre. Each thermometer consists of a loop of bare platinum wire about 5 centimetres long and 1/1000th inch diameter, which is placed in series with a battery of constant potential and a reflecting galvanometer, of short periodic time, the deflection of which is recorded photographically on a геvolving drum. On the same drum the pressure of the gas is recorded. The arrival of the flame at any wire is marked by a sharp rise in its resistance, and the rate of rise, when corrected for the time lag of the wire, gives a measure of the velocity with which the gases about it combine. It is found that with a mixture consisting of one volume of gas and nine of air the flame spreads from the spark in a somewhat irregular manner, but at a rate of roughly 150 centimetres per second. A thermometer placed near the spark shows a sudden rise of temperature to about 1200° C., after which the temperature remains nearly constant until the flame approaches the walls of the vessel. With the rapid rise of pressure which then occurs the adiabatic compression of the burned gas at the centre causes the temperature there to rise to about 1900° C., with the result that the wire of the thermometer generally melts. At a point near the walls the gas is compressed to near the maximum pressure before ignition, and the temperature consequently rises suddenly to 1200° C. or 1300° C., and as there is little subsequent compression there is not much further rise of temperature. Thus, in consequence of the different treatment of the gas at different points in the vessel, differences of temperature of 500° C. exist in the gas at maximum pressure after an explosion of this kind. That such differences must necessarily exist after an explosion even in a vessel impervious to heat does not appear to have been noticed hitherto. These differences are rapidly obliterated by convection currents, but their magnitude at the moment of maximum pressure is such as to make it impossible to obtain an accurate estimate of the specific heat from the pressure record after the manner of Messrs. Mallard and Le Chatelier. The work of these experimenters is not, however, open to the chief objection that has hitherto been urged against it, viz. that combustion was incomplete when they measured the specific heat. The experiments here described show that the combustion at any point is prac