Slanting over a roof, the balloon caught a chimney and tore asunder; but the wreck, also catching, held fast, while the car hung helplessly down a blank wall. In this perilous predicament great coolness and agility alone averted disaster, till firemen were able to come to the rescue. The air ship was damaged beyond repair, but by September 6th another was completed, and on trial appeared to work well until, while travelling at speed, it was brought up and badly strained by the trail rope catching in trees. Early in the next month the young Brazilian was aloft again, with weather conditions entirely in his favour; but again certain minor mishaps prevented his next struggle for the prize, which did not take place till the 19th. On this day a light cross wind was blowing, not sufficient, however, seriously to influence the first stage of the time race, and the outward journey was accomplished with a direct flight in nine minutes. On rounding the tower, however, the wind began to tell prejudicially, and the propeller became deranged. On this, letting his vessel fall off from the wind, Santos Dumont crawled along the framework till he reached the motor, which he succeeded in again setting in working order, though not without a delay of several minutes and some loss of ground. From that point the return journey was accomplished in eight minutes, and the race was, at the time, declared lost by 40 seconds only. The most important and novel feature in the air ships constructed by Santos Dumont was the internal ballonet, inflated automatically by a ventilator, the expedient being designed to preserve the shape of the main balloon itself while meeting the wind. On the whole, it answered well, and took the place of the heavy wire cage used by Zeppelin. M. de Fonvielle, commenting on the achievements of Santos Dumont, wrote:--"It does not appear that he has navigated his balloon against more than very light winds, but in his machinery he has shown such attention to detail that it may reasonably be expected that if he continues to increase his motive power he will, ere long, exceed past performances." Mr. Chanute has a further word to say about the possibility of making balloons navigable. He considers that their size will have to be great to the verge of impracticability and the power of the motor enormous in proportion to its weight. As to flying machines, properly so called, he calculates the best that has been done to be the sustaining of from 27 lbs. to 55 lbs. per horse power by impact upon the air. But Mr. Chanute also argues that the equilibrium is of prime importance, and on this point there could scarcely be a greater authority. No one of living men has given more attention to the problem of "soaring," and it is stated that he has had about a thousand "slides" made by assistants, with different types of machine, and all without the slightest accident.