reductions for variables taken four at a time (table 3)- This combination 

 ranks sixth in importance^ but is separated from the most-important combin- 

 ation by only 1.02^, The suggestion, then, is that the method of data 

 analysis reasonably duplicates the expectable influence of combinations 

 of important variables as they have been suggested by theoretical and 

 laboratory approaches, -when studied by a straight forward least-squares 

 procedure. Then, an important corollary to this inference is that the 

 data set itself is a reasonably faithful representation of the environmental 

 interactions \mder study. 



Exmination of the results (table 3.) for five Xs at a time again 

 shows that the variables found to be _of significance in the laboratory 

 exert the greatest °lo SS reduction on V. It is at this point that the 

 variable of angle of wave approach (XIO) enters into the strongest com- 

 bination. XIO exhibits an increase in its original SS-reduction value of 

 0.65fo, when in the presence (table h) of XI, X3, X5, and XJ. What effect 

 expressing OL and OLq (using Snell's law and Airy wave theory) would have 

 had in bringing the relative ranking of XI, X3, and X5 in table 2 more in 

 line with the powers in expression (l) and (2) is unkown. It may be unim- 

 portant in this set of data, owing to complex wave-refraction patterns 

 that may be present. 



The highest-ranking combinations of six variables at a time include 

 X7, wind velocity offshore, which probably has a significant effect upon 

 wave steepness, X5. Thus the interlocking nature of the variables again 

 enters the picture. On several occasions d\iring the measurement periods 

 wind shifts from onshore to offshore were noticed in connection with 

 passage of cold fronts. Waves were observed being "knocked down," as 

 spray blew from their crests in a seaward direction. Because of the apparent 

 "corroboration" of the method of analysis for the fundamental variables 

 mentioned above, it seems fair to attach significance to the additional- 

 atmospheric variables that turn out to be significant in the least-squares 

 analysis when they combine with these fundamental variables. That is, the 

 indicated importance of variables XT and possibly Xll (table 3) to long- 

 shore-current velocity may be believed to have true physical significance, 

 when in combination with the variables Just mentioned. Once physical 

 significance is attached to these variables it is realistic to enquire into 

 the precise physical relationships involved. As mentioned, for example, 

 strong offshore winds will often rather effectively alter the wave form, 

 and water density enters into fluid di^ag in the longshore- current trough, 

 the amount of sediment entrained, and thereby the velocity of the longshore 

 current. Bruun (1963) and Inman and Bagnold (1963) give excellent reviews 

 of the problems in understanding the generation of longshore currents. 

 Any inferences advanced here are not capable of immediate verification 

 based upon field studies but may serve as a basis for the design of field 

 measurements. 



Because of the strictly artificial introduction of data redundancy 

 in variables X3 and X5 (table 2), a- new summary (table 5) was prepared 



from the computer output. This new table -lists the strongest, combinations 

 when variables X3 and X5 are removed from the analysis. As expected. 



24 



