There were 26 fatal aeroplane accidents between September 17, 1908, and December 3, 1910. Yet in that period there were many thousands of ascents: 1300 were made in one week at the Rheims tournament alone. Of the 26 accidents, 1 was due to a wind squall, 3 to collision, 6 (apparently) to confusion of the aviator, and 12 to mechanical breakage. An analysis of 40 British accidents shows 13 to have been due to engine failures, 10 to alighting on bad ground, 6 to wind gusts, 5 to breakage of the propeller, and 6 to fire and miscellaneous causes. These casualties were not all fatal, although the percentage of fatalities in aeronautic accidents is high. The most serious results were those due to alighting on bad ground; long grass and standing grain being very likely to trip the machine and throw the occupant. French aviators are now strapping themselves to their seats in order to avoid this last danger. [Illustration: THE SANTOS-DUMONT "DEMOISELLE" (From _The Aeroplane_, by Hubbard, Ledeboer and Turner)] Practically all of the accidents occur to those who are flying; but spectators may endanger themselves. During one of the flights of Mauvais at Madrid, in March of the present year, the bystanders rushed through the barriers and out on the field before the machine had well started. A woman was decapitated by the propeller, and four other persons were seriously injured. Nearly all accidents result from one of three causes: bad design, inferior mechanical construction, and the taking of unnecessary risks by the operator. Scientific design at the present writing is perhaps impossible. Our knowledge of the laws of air resistance and sustention is neither accurate nor complete. Much additional study and experiment must be carried on; and some better method of experimenting must be devised than that which sends a man up in the air and waits to see what happens. A thorough scientific analysis will not only make aviation safer, it will aid toward making it commercially important. Further data on propeller proportions and efficiencies, and on strains in the material of screws under aerial conditions, will do much to standardize power plant equipment. The excessive number of engine breakdowns is obviously related to the extremely light weight of the engines employed: better design may actually increase these weights over those customary at present. Great weight reduction is no longer regarded as essential at present speeds in aerial navigation: we have perhaps already gone too far in this respect. Bad workmanship has been more or less unavoidable, since no one has yet had ten years' experience in building aeroplanes. The men who have developed the art have usually been sportsmen rather than mechanics, and only time is necessary to show the impropriety of using "safety pins" and bent wire nails for connections.