cocks and one without, to which our English Board of Trade has to shut its eyes, as the German standpipe, from their point of view, is unsafe and ought not to be allowed.

7. Experiments on the Transfer of Heat through Plates with Variously Arranged Surfaces. By WILLIAM GEORGE WALKER, M.Inst.M.E., A.M.Inst.C.E.

The object of these experiments is to compare the effects on the transfer of heat of variously arranged surfaces projecting from the primary surface of a plate in contact with steam, air or water.

Two cylindrical smooth vessels were constructed of exactly similar dimensions, cut from the same brass tube, 63 inches in length and 2 inches in diameter, fitted with water-tight lids, through which thermometers were inserted. When filled with water no difference was found to exist between their respective powers to absorb or discharge heat. One of the cylindrical vessels was then fitted with copper ribs 5 inches long, inch wide, 012 inch thick, soldered on longitudinally. The ribs were spaced equally round the cylinder and tried as follows:

1. Eight external ribs.

2. Sixteen ribs, eight external and eight internal.

3. Eight internal ribs.

The heating of the two cylinders was performed by suspending them in steam from boiling water. The two cylinders to be compared were filled with water, and placed in steam when their temperatures were 65° Fahr.

The reading of the thermometers, together with the time, was noted simultaneously at every 10 degrees. The time was taken in seconds by a ship's chronometer. When the thermometers became stationary at 210° Fahr. the cylinders were suspended either in air or water and allowed to cool down, the temperatures being noted every 10° Fahr. and the time in seconds.

The difference between the temperatures of the corresponding plain and ribbed cylinders increased from zero and reached a maximum, after a certain time after which they again closed to equal degrees.

The ribbed surfaces increased to a considerable extent the rapidity of transfer of heat either when absorbing heat from steam or discharging it into the air. The effect was not so great when cooled in water. In the externally ribbed cylinder the greatest advance in temperature over the corresponding plain one was 18° Fahr., 33° Fahr., 8° Fahr., when in contact with steam, air, and water respectively. The addition of the internal ribs to the external ones did not produce much effect. With the internal ribs alone, the rapidity of transfer of heat was increased when cooled in water, but practically no effect was perceived when in steam or air.

The external ribs were more effectual in discharging heat to the atmosphere than in absorbing it from steam. This difference may be due to the condensed layer of water which was deposited on the surface.

Coiling wire round was also tried. The temperature of the coiled cylinder fell faster than the corresponding plain one in air, but rose slower in steam-due probably to the condensed layer of water deposited between the wire coils. No difference was noticed when cooled in water. The comparison of rough and smooth surfaces when absorbing heat from steam or discharging it into air or water was also tried.

8. A New Principle of Aerial Navigation. By Lieut. B. BADEN-POWELL.

It has been the constant desire of inventors to devise some means by which we may be able to navigate the atmosphere. Wings, vertical screws, aeroplanes, have all had their advocates, and great hopes were aroused in the balloon. Many proposals were made for steering this aerial buoy, and sails and rudders were applied to the apparatus, until scientists pointed out that these could be of no avail

1

on any apparatus which floated in, and with, one medium. Yet it is this very principle which I wish to advocate, and to state broadly a method by which I believe we might sail through the air, which depends upon three well-established facts. First, a kite retained by a string will ascend when a wind is pressing on its under surface, and will raise a considerable weight. Secondly, in the absence of wind, the same effect may be produced by drawing along the kite through still air. I have myself been lifted by a large kite under such circumstances. Thirdly, by balloon ascents, observations on clouds, mountain records, and especially observations on high places as the Eiffel Tower, it has been found that the wind almost invariably increases in velocity the higher we get, so that the currents of air 1,000 feet up move about three times as fast as those below. It follows from these principles, that if two kites connected by a long string, so arranged that one floats in a current of air blowing at a different rate (or different direction) from that in which the other floats, there will be a reciprocal action, the kite in the lower medium being supported by being drawn along by the kite in the higher stratum, which in its turn is kept aloft by being retarded by the other.

A kite of 1,000 square feet area is capable of supporting a man in a breeze of 10 miles an hour, or when being towed at that rate through calm air. If the wind near the surface of the earth be blowing at this rate (rather below the average), it will usually be travelling at 30 miles per hour at an elevation of 1,000 feet.

With two such kites connected together by a long rope, with a car attached to the rope near the lower kite, if the lower one be drawn along 10 miles an hour faster than the wind, it will support a man, and travel at a rate of 20 miles an hour. But if the upper kite travel at this rate it will be retarded to an extent of 10 miles an hour, and hence the whole apparatus will float along with the wind. In this way we might make a light apparatus for navigating the air. The extent to which it might be steered out of the wind's course, practice alone can determine, but even if this be not much, we still should have an air-ship possessing very many advantages over a balloon, and to which propelling agents could be much more easily applied.

9. Receiver and Condenser Drop. By Professor A. E. ELLIOTT. .

SECTION H.-ANTHROPOLOGY.

PRESIDENT OF THE SECTION.-Professor W. M. FLINDERS PETRIE, D.C.L., L.L.D.

THURSDAY, SEPTEMBER 12.

The PRESIDENT delivered the following Address:—

IN a subject as yet so unmapped as anthropology there is more room for considering different points of view than in a thoroughly organised and limited science. The future structure of this science depends largely on the apprehension of the many different modes of treating it. The time has not yet come when it can be handled as a whole, and therefore at present we may frankly consider various questions from an individual standpoint, without in the least implying that other considerations should not be taken into account. It is only by the free statement, however onesided, of the various separate views of the many subjects involved in such a science, that any comprehensive scheme of its organisation can ever be built up. In remarking, therefore, on some branches at present I shall not attempt a judicial impersonality, but rather try to express some views which have not yet been brought into ordinary currency.

Elaborate definitions of anthropology have been formulated, but such are only too liable to require constant revision as fresh fields of research are added to the domain. In any new country it is far safer to define its limits than to describe all that it includes; and all that can yet be safely done in anthropology is to lay down the 'sphere of influence,' and having secured the boundaries, then develop the resources at leisure. The principal bordering subjects are zoology, metaphysics, economics, literature, and history. So far as these refer to other species, as well as to man, or to individuals rather than to the whole race, they stand apart as subjects; but their relation to the human species as such is essentially a part of anthropology. We must be prepared, therefore, to take anthropology more as the study of man in relation to various and often independent subjects, than as an organic and self-contained science. Human nature is greater than all formulæ; and we may as soon hope to compact its study into a logical structure, as to construct an algebraical equation for predicting its course of thought.

Two of the commonest and most delightfully elastic words in the subject may be looked at once more-'race' and 'civilisation.' The definition of the nature of race is the most requisite element for any clear ideas about man. Our present conception of the word has been modified recently more than may be supposed by our realising the antiquity of the species. When only a few thousand years had to be dealt with nothing seemed easier or more satisfactory than to map out races on the assumption that so many million people were descended from one ancestor and so many from another. Mixed races were glibly separated from pure races,

and all humanity was partitioned off into well-defined divisions. But when the long ages of man's history and the incessant mixtures that have taken place during the brief end of it that is recorded come to be realised, the meaning of 'race must be wholly revised. And this revision has not yet taken effect on the modes of thought, though it may have demanded the assent of the judgment. The only meaning that a 'race' can have is a group of persons whose type has become unified by their rate of assimilation and affection by their conditions exceeding the rate of change produced by foreign elements. If the rate of mixture exceeds that of assimilation, then the people are a mixed race, or a mere agglomeration, like the population of the United States. The greatest problems awaiting solution are the conditions and rate of assimilation of races-namely, what period and kind of life is needed for climatic and other causes to have effect on the constitution and structure, what are the causes of permanence of type, and what relative powers of absorption one race has over another. Until these problems are reduced to something that can be reasonably estimated we shall only grope in the dark as to all racial questions.

How, then, can these essential problems be attacked? Not by any study of the lower races, but rather by means of those whose history is best recorded. The great mode of isolation on which we can work is religious difference, and oppressed religious minorities are the finest anthropological material. The first question is given a mixture of various races in approximately known proportions, isolated, and kept under uniform conditions, how soon does uniformity of type prevail? or what proportions of diversity will be found after a given number of generations? A perfect case of this awaits study in the Copts, who have by monogamy and the fanaticism of a hostile majority been rigorously isolated during 1,200 years from any appreciable admixture, and who before this settling time were compounded of eight or ten different races, whose nature and extent of combination can be tolerably appraised. A thorough study of the present people and their forefathers, whose tombs of every age provide abundant material for examination, promises to clear up one of the greatest questions-the effect of climate and conditions on assimilating mixed peoples. The other great problem is, How far can a type resist changes of conditions; provided it be not mixed in blood, so as to disturb its equilibrium of constitution? This is to be answered by the Jews and the Parsis. As with the Copts, an oppressed religious minority has no chance of mixture, as all mixed marriages are abhorrent to its exclusiveness, and are at once swept into the hostile majority. The study is, however, far more difficult owing to the absence of such good conditions of the preservation of material. But nothing could throw so much light on this as an excavation of some Jewish cemeteries of a thousand years or so ago in various European countries, and comparison of the skeletons with the proportions of the Jews now living. The countries least affected by the various proscriptions and emigrations of the race would be the proper ground for inquiry. When these studies have been made we shall begin to understand what the constants of a race really are.

We will now look at another word which is incessantly used-'civilisation." Many definitions of this have been made, from that of the Turk drinking champagne, who remarked about it that after all, civilisation is very nice,' up to the most elaborate combinations of art and science. It is no doubt very comfortable to have a word which only implies a tendency, and to which everyone can assign his own value; but the day of reckoning comes, when it is brought into arguments as a term. Civilisation really means simply the art of living in a community, or the checks and counterchecks, the division of labour, and the conveniences that arise from common action when a group of men live in close relation to each other. This will perhaps be objected to as including all—or nearly all-mankind in its scope. Quite true; all civilisation is relative and not absolute.

We shall avoid much confusion if we distinguish high and low types of civilisation, and also perfect and imperfect civilisation. Like organisms we may have a low type of civilisation very perfect in its structure, capable of endless continuance, and of great shocks without much injury. Such are some of the civilisations 1895.

3 G

of the African races who have great orderliness and cleanliness of arrangements, and are capable of active recuperation after warfare, without any internal elements of instability. Again, some low types are very imperfect, and can exist only by destruction of others, while any severe shock destroys their polity; the governments which only exist by raids and plunder, such as that of the Zulus, illustrate this. Turning to high types of civilisation we may see them perfect or imperfect. Countries of financial stability, not undergoing any rapid organic changes, are the more perfect in type; while those deeply in debt and in continual revolution have but imperfect civilisation, of however high a type it may be. With these distinctions before us,-that all civilisation is a question of degree, that there are types of all variety, from the highest complexity to the lowest simplicity, and of all degrees of perfection, or stability and completion, in any given level of complexity-with these distinctions some of the vagueness of verbal usage may perhaps be avoided.

Turning now from words to things, we may perhaps see some ground for further consideration in even one of the best elaborated departments.

In the much-vexed question of skull measurements, the paucity of clearly defined racial characteristics may make us look more closely as to whether we are working on an analytic or an empirical method. In any physical problem the first consideration is the disentangling of variables, and isolation of each factor for separate study. In skulls, however, the main measures are the length, which is compounded of half a dozen elements of growth, and the breadth and height, each the resultant of at least three elements. Two skulls may differ altogether in their proportions and forms, and yet yield identical measures in length, breadth, and height. How can any but empirical results be evolved from such a system of measurement alone?

A departure from this mechanical method has appeared in Italy last year by Professor Sergi. He proposes to classify skulls by their forms,-ellipsoid, pentagonoid, rhomboid, ovoid, &c. This, at least, takes account of the obvious differences which the numerous measurements wholly ignore. And if skulls were crystals, divisible into homogeneous classes, such a system would work; only, like all organic objects, they vary by infinite gradation.

What then lies behind this variety of form? The variety of action in the separate elements of growth. Sergi's ellipsoid type means slight curvatures, with plenty of frontal growth. His pentagonoid means sharper curvatures. His rhomboid means sharp curvatures with small frontal growth. And so in each class, we have not to deal with a geometrical figure, but with varying curvatures of the centre of each plate of the skull, and varying extent of growth from the

centres.

The organic definition of a skull must depend on the statement of the energy and direction of each of the separate elements out of which it is built. The protuberances or eminences are the first point to notice. They record in their curves the size of the head when it attained rigidity in the centres of growth. Every person bears the fixed outline of parts of his infant skull. Little, if any, modification is made in the sharpness of the curves between infancy and full growth; perhaps the only change is made in course of the thickening of the skull. Hence the minimum radius of curvature of each plate of the skull is a most radical measurement, as implying early or late final ossification. In higher races finely rounded skulls with slight curvatures are more often found; and this agrees with the deferred fixation of the skull pointed out by the greater frequency of visible sutures remaining, both effects being probably due to the need of accommodating a more continued growth of the brain. The length of growth of each plate from its centre in different directions regulates the entire form of the skull. The maximum breadth being far back implies that the parietals grow mostly toward the frontal or vice versa. The top being ridged means that the parietals grow conical and not spherically curved, and hence meet at an angle.

It seems, therefore, that looking at the question as a physical problem, we are far more likely to detect racial peculiarities in the separate data of the period of fixation of the skull, and of the amount of growth in different directions, than