ductor d is located. On the top of tube f a flexible rail b is secured, and this connects with studs c, which are within the tube, as clearly shown in the drawing, and so situated that they may be forced down into contact with d. Normally these studs are separated from d, but when the car comes along, the wheel a, mounted upon the end of plow P, flattens the tube f and thus forces one or more of the studs c into contact with d. The distance between the studs c is such that at least two will always be in contact with d, thus insuring a continuous electrical connection with the motors so long as the plow is depressed.
The first impression upon looking at this design would be that it is entirely free from objections; for if we assume that the tube f is made of rubber, we can see it in our mind's eye springing up after the plow passes by and thus separating the contacts c from d, and at the same time yielding freely to the pressure of the wheel a. All of this is true, but rubber is not very durable when under such exposed conditions, and to maintain a length of several miles of it in a perfect state for even two or three years could not reasonably be expected; and if it became necessary to renew the tube oftener than this the cost of maintenance would be entirely too great. There is another objection, however, which is more serious, and that is that the conduit will gradually fill up with dirt, and this pressing against the rubber tube would force it out of shape, and thus cause the contacts c to bear permanently upon d, or else to become so far displaced that they would not touch it when depressed by the plow.
As the rubber tube can not be depended upon, inventors have sought to improve the construction by using sheet steel and making the tube flatter and much wider, so that a section of it would present an outline much resembling an elliptic carriage spring. Such a construction will meet the requirements as to strength and the retention of the contacts c in their proper position; but steel expands when warm and contracts when cooled, therefore a long tube would be stretched so much in winter that it might pull apart, while in summer it would be compressed and tend to buckle up and thus be forced out of place. These difficulties can be overcome by providing expansion joints at suitable intervals, so that they are not necessarily proof of the impracticability of devices based upon the principles involved in this design; they simply serve to forcibly bring to mind the fact that the path of the inventor of underground systems is not strewn with roses, no matter in what direction he may turn to find a solution of the problem.
The object in the designs Figs. 25 and 26 is to shield the conductor so that it will remain dry should the conduit be filled or par-
