Although the estimates of total pool area 

 were adequate, estimates of pool area by 

 quality strata naturally would be less precise 

 because of the greater amount of variation 



between samples. Consequently, such esti- 

 mates could be considered as only being in- 

 dicative of the amount of pool area by quality 

 classes. 



DISCUSSION 



Although the procedures tested in this 

 study provided good results, better results 

 undoubtedly could be achieved by adopting 

 several modifications. 



We recommend that a minimum of five 

 transects be measured at each sample point 

 instead of a single transect. This would have 

 these advantages: (1) reduced variation be- 

 tween samples, particularly in the quantity 

 of pool area; and (2) reduced number of 

 samples required to adequately describe a 

 drainage. Such reduction in number of sam- 

 ples means that the distance between samples 

 along the channels could be increased from 

 one-fourth mile to as much as 1-mile intervals. 



Regardless of the length of the stream 

 being surveyed, at least In samples — prefer- 

 ably 30 — are required. Thus, the actual 

 interval between sampling points depends on 

 the length of the stream. In general, the 1- 

 mile interval would be adequate for streams 

 that are 15-30 miles long. Closer spacing 

 would be necessary on shorter streams to ob- 

 tain sufficient samples for maintaining accept- 

 able limits of precision. 



Procedures for measuring and recording 

 stream characteristics would remain un- 

 changed, except that stream gradients would 

 have to be measured only for the first and 

 fifth transects rather than for all five. When 

 calculating the average value of any specific 

 characteristic for a channel, the estimate 



would be based on five transects rather than 

 on one. 



It usually took about 15 minutes to meas- 

 ure a single transect and another 15 minutes 

 to walk one-fourth mile to the next sample 

 point. This meant it took approximately 2 

 hours to measure the four transects in a mile 

 of stream. We estimate that a five-transect 

 sample could be measured in about 45 minutes 

 to an hour. Since the travel time on foot 

 would be the same, time requirements per 

 mile of stream could be reduced slightly. 



The time saved would be greater on 

 streams that are paralleled by nearby roads 

 that could be used for auto travel between 

 sampling points located at 1-mile intervals. 

 Auto travel between sample points located at 

 14 -mile intervals proved impractical on the 

 streams measured in this study because the 

 roads usually were located so that the round- 

 trip walk from auto to sample point was 

 almost as far as the one-fourth mile from one 

 sample point to the next. 



Procedures also might need to be modified 

 for larger streams. The streams examined in 

 this study seldom exceeded one hundred feet 

 between banks, and data usually could be 

 obtained by fieldworkers using waders. For 

 larger streams, however, it may be necessary 

 to use boats. Sampling of bottom composi- 

 tion also would require changed procedures 

 on larger streams. 



12 



