22 BULLETIN 289, U. S. DEPARTMENT OF AGRICULTURE. 



(E) Indian palmetto fiber, coarser than any of the others, but somewhat brittle. 

 A portion of this brush was modified by cutting out sections of the bristles to make 

 a more uneven surface. 



(F) Ordinary bristle hairbrushes. 



(G) Same as brush E, but with about half of the bristles clipped out. 

 (H) Same as brush C, but with about half of the bristles clipped out. 



The idea of utilizing pairs of brushes with different types of fibers 

 was to determine, if possible, if any of them would give sufficient 

 promise to warrant an application of the particular type of brush 

 to a mechanically operated machine that would imitate the action 

 of the small brush when operated by hand. For this reason no 

 brushes were used which could not be duplicated on a machine for 

 operation on a field scale. 



The work with the handnyperated brushes was done principally at 

 La Fayette, Ind., and Ames, Iowa. In some experiments the heads 

 were manipulated with the brushes at different times of day, while in 

 other experiments different numbers of treatments were given the 

 heads at varying intervals. The direction from which the heads were 

 struck also varied in certain experiments, some being given vertical 

 strokes and others lateral strokes. It was thought that cross-pollina- 

 tion might be brought about by the vertical stroke, which apparently 

 would enable some of the brush bristles to spring the keels and convey 

 pollen from one flower to another. It was also thought that if the 

 flowers were self-fertile the lateral strokes would accomplish this. 

 The representative tables that follow indicate the principal features 

 brought out by this series of experiments. 



relative efficiency of brushes when the clover heads were struck 



Horizontally. 



In a clover field 2\ miles east of La Fayette, Ind., 26 square-rod 

 plats were laid off for this experiment in 1911 (Table VI). All of 

 the heads in bloom at the time the plats were marked off were re- 

 moved. 



Plat 1 was left as a check, no brushes being used on it. Plat 2 was 

 worked with brushes A, one brush being taken in each hand and the 

 blossoms struck between the brushes by a quick movement of the wrists. 

 When in full bloom the heads received one treatment, the operator 

 going only in one direction across the plat. Brushes B, C, D, E, and 

 F were used on plats 3, 4, 5, 6, and 7, respectively, in the same manner 

 as for brushes A on plat 2. Plat 8 was treated by going one way 

 across it at right angles to the first way, thus giving each blossom two 

 treatments, the strokes of the two treatments thus being at right 

 angles to each other. Brushes A were used. Plats 9, 10, 11, 12, and 

 13 each received treatment similar to plat 8, but with brushes B, C, 

 D, E, and F, respectively. Plats 14 and 15 were left as checks. 

 Plats 16. 17, IS, 19, 20, and 21 each received two treatments with 



