This paper describes an investigation conducted at the request of the Aeronautical Research Committee, who made a grant to one of us for the purpose. It is submitted by permission as an essay to compete for the R.38 Memorial Prize.

No essential alteration has been made in the form of presentation adopted in our preliminary reports, but to meet the needs of the general reader we have added the introductory paragraphs which follow. These explain briefly the nature of the problem which we have set ourselves, the range of our solution, and the bearing of our results on practical questions which confront the designer of a rigid airship hull.

The establishment of contact between aircraft in flight for the purpose of refuelling, has until recently been regarded as a somewhat hazardous procedure employed mainly during attempts to create endurance records.

It would appear, however, that if such contact between aircraft could be established safely and as a routine operation, considerable advantages could be obtained by the employment of such a scheme.

From the point of view of commercial aviation, not only would it make possible long distance non-stop flights over regular air routes, but it would enable a large increase in pay load to be carried, since the bulk of the aircraft's consumable load, i.e., fuel, would not be taken on board until the aircraft had left the ground and climbed to a reasonable height.

In my article Wing-Beats, which appeared in this Journal, in the issue of October, 1930, I recorded a series of experiments demonstrating the strength and lightness of a spring formed by twisting skeins of natural silk.

My hope is that silk torses of this kind will fulfil, in human flight, the postural function of a bird's pectoral muscles, leaving to human effort the phasic function of these muscles, that is to say, the activity induced by rapid nerve impulses from the brain.

The ratio that may subsist between these two functions cannot, I think, be determined otherwise than by full-scale flight experiments, but the experiments already recorded give reasonable ground for the belief that the requisite human phasic output will be small in relation to the total weight movement.