This layout gives one-and-a-half minute service in each direction.

The preceding method of handling the cages by means of travelers was not designed instanter, but is the result of a gradual development through a series of tentative studies. The first idea that occurred to the author was to carry the cages by the suspension system around a loop of large diameter at each end of the route, but this necessitated a very awkward detail, troublesome to operate, consisting of a pair of movable platforms to permit cars and other vehicles to enter or leave the cage at the ends. This alone was sufficient reason for condemning the method; but there was a still-more-important one, viz., that a breakdown of one of the traveling cages would tie up the whole line. This contingency might have been provided for by having outside of each loop a traveler leading to a "hospital" pocket; but that detail also would have been awkward and troublesome to operate, hence the loop-system project was soon abandoned.

Next came the idea of using switches at the approaches, but it was evident at once that overhead ones were impracticable on account of the necessity for cutting through the supporting girders, hence they would have to be located beneath the cage, which would then require trucks below as well as above, and the support would have to be changed at very short intervals between the upper and the lower systems. This method, too, was awkward, requiring both time and an additional expenditure of power to make the change, consequently it, also, was soon abandoned.

Next came the scheme of operating the two tracks as entirely independent units, but adjusting the times for starting so as to have comparatively regular intervals between cages in each direction. This scheme involved the idea of the lateral travelers running on transverse tracks below, with the cage still supported from above by longitudinal tracks which would be a continuation of those on the span, the interval between being so small as readily to be jumped by the wheels.

The layout for this method of operation is shown in Fig. 32b, from which it is seen that near the outer end of each trestle-approach there is to be a pair of independent, double-chambered travelers running on transverse tracks, the extreme motion of each traveler being about 25 feet. These travelers are entirely disconnected from the top of the trestle, but lie up very close to it. Beyond each traveler is a stationary pocket (large enough to receive a cage) braced at the sides, but open at the river end for the full height, and open at the shore end near the bottom high enough for the ingress and egress of vehicles and pedestrians.

The modus operandi is as follows: As shown at the right-hand side of Fig. 32b, looking from the river, the traveler is at its outer position, all three of the cage-receptacles being vacant. Cage 1 arrives and runs into the inner chamber of the traveler, which is then moved inward so as to allow the contents of the cage to be unloaded and a new load to be taken on. While this  is  occurring,  Cage 2  arrives and passes across the outer chamber of the traveler  to  the  pocket  beyond,  and  from  there  the  unloading