for the pools. There is one notable exception, 

 however, Winter Creek, which is near Prairie du 

 Chien in pool IQ produced fishing in each of the 

 two winters which averaged from 3- to 10 times 

 better than the fishing at the various other local- 

 ities. This probably was due to an underground 

 spring, which keeps a fair sized area free from 

 ice throughout the winter. Apparently the blue- 

 gills are attracted to and induced to feed in this 

 open water area and hence are easily caught. 



Table 17 gives the total numbers of 

 separate localities which were censused in each 

 of the 4 seasons. These numbers are substan- 

 tial, especially for the summer of 1945 and the 

 winter of 1944-45. Many of the tabulated 

 "localities" were of minor significance in their 

 contributions to general averages, since for many 

 of them less than 100 or 200 fisherman -hours 

 were recorded for an entire season. The com- 

 plete list of locality names is not given in this 

 report . 



In table 17 is listed the numbers of local- 

 ities, for each of the seasons, with a total of 

 more than 300 hours fishing per locality per 

 season. Fishing success for the localities is 

 given in terms of the range and the mean. The 

 ranges were broad and the deviations large, as 

 shown by the fact that V (= \QOS/ M) runs as high 

 as 44 for the winter of 1944-45, and 76 for the 

 summer of 1945. 



These figures are given in terms of total 

 fish per hour, regardless of kinds and sizes of 

 fish. Some of the qualitative aspects (species 

 composition of the catch) for certain of the local- 

 ities have been discussed above. 



Variation throughout the season. Day-to-day 

 fluctuations in the average fishing success (total, 

 all pools) for the summer 1945 are shown in 

 figure 1. As explained above, each field worker 

 usually found it impossible to cover his territory 

 completely on any given day . Therefore, his 

 returns for successive days often represented 

 separate localities or portions of his territory. 

 This factor probably influenced the apparent 

 daily averages for the total area, causing them 

 to vary from day to day more than would have 

 been the case had it been possible to cover all 

 localities every day. The daily averages for the 



entire season (total all pools) showed a range of 

 from 0.20 to 1.23 fish per hour, with a mean of 

 0.54 and a standard deviation of 0.21 . This 

 gives a fairly high coefficient of variation, V = 39. 



To alleviate this effect of daily fluctuation, 

 the graph of figure 1 is constructed with moving 

 averages of fives. Even so, the line still has a 

 jagged appearance. It is difficult to say just how 

 much of this fluctuation from one day to the next 

 is real and how much is due to the circumstance 

 explained above. Hansen (1942) found that for 

 Lake Chautanqua, the periods of good fishing 

 tended to be short, with ateep-sided curves. In 

 three Indian lakes, as reported by Ricker (1945), 

 the week-to-week variations throughout the sum- 

 mer were very irregular, and were different for 

 different species. Some of the variations cover- 

 ing a week or more probably are authentic and 

 are tied in with physical factors, although these 

 may be but vaguely evident . 



In general, the fishing success ratio from 

 about mid-July on was higher and also subject to 

 more fluctuation than it had been during the early 

 part of the summer . This holds true for most of 

 the entire area, but comes about especially be- 

 cause of the pronounced tendency in this direction 

 of the more down-river parts of the area (i.e., 

 pool 10 and pool 11). This is brought out in 

 figure 2, which shows comparative graphs for 

 pool 4L, pool 10, and the total (all pools) . In 

 this figure, weekly rather than daily averages 

 are used, of necessity, since for any given pool 

 the effect of working various days is consider- 

 able . During any given week, however, an 

 entire pool was fairly well covered. 



The graph for pool 10 follows closely the 

 contours of the graph for the total (all pools) . 

 The existence of a definite mathematical correla- 

 tionship will be pointed out below . On the other 

 hand, the line for pool 4-L is much more steady 

 and level, showing only a minor peak during the 

 mid-July period. These two pools, being some 

 distance apart and very different physiographic - 

 ally, were subject to different sets of physical 

 conditions throughout the season. Furthermore, 

 the fisheries of the 2 pools were of different 

 natures, the catch of pool 4-L being mostly wall- 

 eye and sauger, while that in pool 10 ran largely 

 to bluegill, crappie, and bass . This difference 



27 



