Papanikolaou and Nowacki [18]. The computed results seem to be good except 

 for the second-order amplitude F2i. The discrepancy In the second-order force 

 amplitude is probably due to insufficient accuracy in the computed nonlinear 

 force. Since first-order contributions to the force are dominant, the 



TABLE 1 - FIRST- AND SECOND-ORDER FORCE COEFFICIENTS 





Lee 



Papanikolaou 



and 



Nowacki 



Computed 



Nonlinear 



Results 



Fl 



0.69 



0.68 



0.67 



«1 



350° 



357° 



357° 



F2O 



-0.2 



-0.19 



-0.23 



F2I 



0.62 



0.68 



0.82 



62 



5° 



107° 



108° 



relative accuracy of the remaining force when the first-order predictions 

 are subtracted is small. The error is compounded since values calculated from 

 the remainder are divided by the square of the perturbation parameter, a 

 small number, to get F20 and F21 . The computed second-order phase angle 

 is, however, excellent. 



22 



