ever; the bird faced into the air flow for only an instant, then turned to fly 

 in another direction, usually downwind. It appeared that the impetus of 

 the upthrust into the air was a reference by which the bird made its orien- 

 tation into the wind, for it was only at this starting instant that the hooded 

 birds seemed truly sensible of a steady flow of air. 



It was typical of the hooded bird to carry its flight downwind from the 

 starting place, and exceptions were few (Table 1).* In slow-flying birds, 

 like English Sparrows and Yellow-headed Blackbirds, this downwind drift 

 was usually in patterns of circles. After facing the wind for a moment, the 

 bird turned to swing around in a circle, and each complete turn was a leg 

 in the downwind shift, as indicated in Figure 8, A, C. The direction of this 

 circling was not constant ; on any given day some flew clockwise, others coun- 

 terclockwise. Several times when we flew the same bird twice, it circled to the 

 right on the first flight and to the left on the second. 



In a breeze of less than 5 m.p.h. the circles were well formed and the 

 downwind progress was slow. With an increased wind velocity the flight 

 pattern showed a series of loops (Figure 8, D, E, F) rather than full circles, 

 and the downwind progress was more rapid. In flying these loops, the hooded 

 sparrow or blackbird seemed to hang in the air a moment on the upwind turn, 

 its ground speed reduced or, if the speed of the wind was the same as the 

 air speed of the bird, stopped completely. After a moment of hovering into 

 the wind, the bird continued with its circle. When the wind velocity was 

 greater than the air speed of the bird, it traveled backward over the ground 

 during the moment it faced the wind. In heavy winds the pattern of the 



* Tables 1 and 2 include only the hooded flights which were made in 1950, '51, and '52. 

 From this point the discussion is based entirely on the results of these flights, for which com- 

 plete data were recorded. Although notes on behavior were made in all 249 flights, imperfect 

 recording techniques made it impossible to include the earlier experiments in Tables 1 and 2. 



65 



