WE are so much accustomed to speak of the perfection of the track and equipment of our leading railroads that the occurrence of such a disaster as happened last week at Bridgeport to the Washington express flls us with dismay and raises a very grave doubt as to whether we have even yet discovered how to make the modern high-speed train a perfectly safe means of travel. Broadly speaking, railroad accidents are due to failure of either the physical or the human element --the roadbed and rolling stock or the operating employee. The Bridgeport accident, according to the statement given out by the ofcials of the road, was due immediately to the failure of the engineer in charge of the train; remotely, however, it must be regarded as due to track conditions, which, had the been other than they were, would have prevented the accident in spite of the carelessness of the employee. The line of the New Haven Railroad at the scene of the accident is a four-track road, the outer tracks of which are mainly reserved for local or suburban service, and the two inside tracks for express service. At the stations the express trains, unless express platforms are provided, make use of the local. platforms, crossing over for this purpose from the express to the local tracks at a point some distance back from the station. It is customary for the expresses to slow down on approaching the switch of the cross-over, and it seems that, in the case in question, the rules of the New Haven road call for a reduction of speed to fifteen miles an hour. The officials of the road have stated that the engineer ran his heavy express train over the switch at a speed that was nearly sixty miles an hour. Now, at any such speed as this, the inertia of the engine was too great for the flanges of the wheels to do their work effectively in holding the train on the track. It mounted the rails and plunged down the embankment, with the result that some fourteen people were killed and half a hundred wounded. Unquestionably, in this day of high-class track and rolling stock, the most prolific source of accidents on railroads is the failure of the human element. For some reason which will never be explained, since the man is dead, the engineer either failed to observe or to obey both the signals and the speed regulation. It was hot weather and he had gone through a heavy day's work-possibly he was in a state of physical weariness which caused a temporary mental lapse. Whatever the explanation, the fact remains that we have here one more railroad disaster attributable to human failure; and the question immediately presents itself, “Would it not be possible to arrange this matter of crossovers or turn-outs for fast trains, so that the safety of the train would be automatically taken care of, and not be left dependent on the mental and physical condition of the engineer?" The answer to this question is that it would be entirely possible, in the great majority of cases, to so construct these cross-overs that even if regula- tions as to slow running were overlooked or disobeyed, an express train would run through from one track to the other without any serious risk of derailment. This could be done by making the cross-over extend over a greater distance, measured along the axis of the roadbed, and making the angle of the frogs, switches, etc., as small as is compatible with structural requirements. It would be entirely possible to lay out the track with switches and curves so easy that if a heavy express train disobeyed its orders (say for a 30-mile speed at the cross-over), and swept over it at a speed of 60 miles an hour, it could do so without any grave risk of derailment. N ow we understand that at least one of our leading railroads has already adopted this plan. Its cross-overs are so gradual and the length of the cross-over between switches is so great, that the trains are swung from one track to the other, even when they are running at high speed, at the will of the signal man, and without the imposition of any speed reduction upon the engineer. These are the ideal conditions; for construction of this character eliminates entirely a risk of operation, whose magnitude is only appreciated in the presence of such a frightful disaster as that which recently occurred near Bridgeport. "Ve commend this most important subject to the joint consideration of the Interstate Railroad Commission, and of the maintenance of way engineers of our railroads. The lengthening out of cross-overs would be one more step in the elimination of the human element as a predisposing cause for accidents, and therefore it would be consistent with the policy which has brought about the introduction of the automatic block signal system and similar safety appliances for the protection of life in railroad travel. Causes of Aviation Accidents ACCORDING to the conclusions of Lieut.-Col. Bouttieaux, who has just prepared a careful report upon the princip31 causes of serious aviation accidents, these may be divided into four classes, due respectively to faults of construction, mistakes of the pilot, state of the weather, and fnally, either to the fault of the public or to one of a variety of special or ill-determined factors. Up to the frst of the present year, there is a record of 31 fatal accidents, costing the lives of 34 persons. Of these, twice as many took place in 1910 as in 1909; but if we take into account the fact that the number of aviators had been quintupled, the proportion descends to 40 per cent of the previous year. Seventeen of these accidents were due to bad construction, 2 to atmospheric perturbations, 9 to faults of the pilot, and 3 to other causes. The classifcation of simple accidents, i. e., those without tragic results, gives in the above-mentioned classes the fgures 26, 27, 33, and 31. Hence, in the total, 29 per cent of these accidents are imputable to the construction and 27 to the piloting of the machines. The following summary of the accidents per year is illuminating: First, imperfect construction was responsible for 14 accidents, 1 of which was fatal, in 1909, and in 1910, for 29 accidents, 16 of which were fatal. Second, errors of pilot caused 21 accidents (1 fatal) in 1909, and 21 accidents in 1910, of which 8 were fatal. Third, atmospheric disturbances caused 4 accidents (none fatal) in 1909, and 25 (2 fatal) in 1910. Fourth, imprudences of spectators or aviators and other special or ill-determined causes were responsible for 8 accidents (1 fatal), in 1909, and for 26 accidents in 1910, of which 2 were fatal. The total number of accidents in 1909 was 47, and in 1910 there were 101 accidents. The increase of accidents due to atmospheric conditions is explained largely by the fact that in 1909 aviators dared not risk their lives in all sorts of weather, as they began to do later, while the greatly increased number of aviation meets had as a corollary an increase in the number of accidents of class 4, caused by faults of spectators or collisions in a restricted landing place. Faults of construction may be divided into those which may be ascribed to the principle of the machine, and those due to imperfect execution of a plan theoretically perfect. The proportion of the frst to the second is as 2 to 43, which does honor to the conception of the aeroplane. Among accidents of construction, those involving the wings amount to 18, and they frequently were due to a rupture at the line of union.with the fuselage, especially in monoplanes. The results are among the most terrible, with 11 cases fatal, and the others very serious. Those asclibable to the motor or controls were 17 in number, of which 5 were fatal. The propeller entered into 6 accidents, in 5 of which it was of metal. Errors of pilotage brought abput 42 accidents, of which 9 were fatal. Among tlfese, faulty turns entered into 24, of which 3 were fatal, while the manner of landing occasioned 8, of which 5 were fatal. Too sudden an ascent, causing a 10ss of sustaining power to the machine, in consequence of the insuf-cient power of the motor, caused 9 accidents, with 1 fatality. ' Of reciprocal injuries between aviators, 5 resulted from too close a proximity of passage and the resulting air “wash." There is nothing to show whether biplane or monoplane is the more subject to accidents, for if more fell to the share of the former, th:t is due, for the moment at least, to the circumstancf; of their majority; the percentage was about the SLlne. How may these causes of acciden be overcome? The author of the report advises, in ;he frst place, the use by pilots of helmets, belts, antI other safety devices in the shape of garments, wciich will be efcacious in cases of moderate fall. for machines he recommends shock absorbers and especially a lightly-constructed prow, the breaking of which would serve to deaden the shock. It wot:ld, in fact, be desirable that special experiments sho·tld be made with reference to the localization of a :weak point upon which the initial efect of a shock would be expended, just as in every electric circuit a fuse is introduced at some point where will be localized the efects of a chance excess of voltage. As for parachutes, their tests have thus far been too few to be commented upon. Dangers from faulty construction may be minimized by demanding a certifcate of navigability and enforcing certain requirements of construction with reference to the choice of mater;,ls and the strengthening of the controls. For the pilot the addition of automatirt stabilizing appara tus would be desirable ; but these—must be practical and efcacious, which is at, PJesent oIi['Ti hope. Equally useful would be insttuments indicating loss of speed, exaggerated ip:cidence, lateral sliding toward the center in turning; etc. The education of the pilot, irproving progressively, on his part, will tend t<.1 eliminate risks of aviation, while greater difficulty.n obtaining a pilot's license will, to some extent, limit accidents to the period of apprenticeship. Cauualties due to meteorological conditions are mare-! aifficult to avoid, if “not impossible. The best remedy would be a deeper knowledge of meteorology and topography, and the making of appropriate topographical maps will contribute to this end. The edqcation of crowds, as well as fewer aviation meets, vll diminish the risks occasioned by them. The conclusions of the report are very encouraging. Bouttieaux proves that the increase of accidents, so alarming at the frst view, arises from the great progress made by aviation, and this in spite of the indisputably rudimentary character of several of the machines. The improvement: of these, whose era is just dawning, will render them steadily safer and more manageable, till a reason!3ble degree of safety is assured. The Albedo of Clouds T. HE albedo of clouds varitfes between wide limits , but w as fo rmerly aumed to average about 0.75; i. e., the up ps{r surface of the clouds was supposed to refect about 75 per cent of the incident sunlight. The albedo of white paper is 0.70; of new-fallen snow 0.78. The frst attempts to measure accurately the albedo of a layer of clouds seen beneath an observer posted on the top of a mountain were those made by Abbot and Fowle on Mount Wilson in 1906, and gave an average of 0.65) but this result was later found to be doubtful, owing to errors in reduction. Within the past few months this problem, which has such important bearings upon the physics of the earth's atmosphere, has been taken up in Germany by Messrs. Stuchtey and Wegener, who made numerous measurements with a specially constructed albedometer in the course of several balloon voyages. They found the following values, which have been corrected by eliminating the general radiation of the sky, and refer only to the proportion of direct sunlight refected: Lower stratus clouds, 0.54; higher stratus, 0.76; cumulus, 0.67. They also measured the albedo of the earth's surface as seen from altitudes between 600 and 1,650 meters. The albedo of open felds was found to average 0.15; of woods, 0.06.