414 JOURNAL OF THE WASHINGTON ACADEMY OF SCIENCES VOL. U, NO. 17 



where U takes the form 



^ yk 



b is the width of fin where good thermal contact- occurs along one 

 edge of the fin only, and is half the fin width where good thermal 

 contact is made by both parallel edges. The fin thickness is y and 

 the thermal conductivity k, the quantity q being the surface heat 

 dissipation coefficient in units consistent with k and the dimensions 

 b and y. For the combinations of values usually occurring in radia- 

 tor practice, the effectiveness exceeds 70 per cent and quite commonly 

 averages 85 per cent to 95 per cent. The formula was applied to 

 calculate optimum dimensions for fins under various conditions, 

 the results being tabulated. The equations of thermal conduction 

 are also used to compute a table of temperature drop through water- 

 tube walls of various thickness and materials. 



A considerable amount of detail not suited to abstracting is to be 

 found in short notes on such topics as effect of yaw on performance, 

 slipstream mountings, wing radiators, shuttering, radiator require- 

 ments at altitude and methods of computing performance at alti- 

 tude from test data obtained at ground level. A brief treatment is 

 included of the subjects of partially streamlined casings to enclose a 

 radiator, and of nose mounting. 



From the point of view of an airplane designer, it is desirable to 

 specify radiator performance in terms of flying speed of the plane. 

 In general this cannot be done directly because too many factors are 

 involved in the relation between flying speed and the air flow through 

 the radiator core. It becomes necessary, therefore, for describing 

 characteristics of the radiator, to adopt air flow through the core as 

 the prime variable, there being no other quantity or entity available 

 for the purpose. The ease or difficulty of applying such a description 

 of performance to plane design rests then with the difficulty in esti- 

 mating the probable air flow through the core for given flying speed. 

 For a radiator mounted in an unobstructed position the difficulties 

 are inappreciable, the estimation process is replaceable by actual 

 measurement and the results are obtainable with satisfactory accuracy. 

 For a nose mounting, or other obstructed position too many factors enter 

 into the determination of the air flow to make generalizations possible. 

 Consequently, the authors are unable to include in the paper material 



^ Soldered or integral rnetal. Mere contact, unsoldered, of two pieces o[ metal is ther- 

 mally a discontinuity. 



