Furthermore, when a large area is under con- 

 sideration, water temperature from one place to 

 another may vary considerably. Surface temper- 

 atures are under the influence of atmospheric 

 conditions and may change from time to time 

 throughout a given period. Probably for these 

 reasons few authors have attempted to demon- 

 strate correlations between water temperature 

 and recorded catch . 



In the summer of 1945, water tempera- 

 ture was recorded somewhat irregularly at 

 several places, all in pool 10. The most com- 

 plete of these sets of data is that for the locality 

 "Railroad Bridge" at Prairie du Chien. Here the 

 water temperature was measured almost every 

 day from about June 10 to the end of August. 

 These figures are averaged week by week and 

 are presented in table 21 . To a reasonable ex- 

 tent, these readings may be used as representa- 

 tive of the water temperature changes throughout 

 most of the entire area. It has been observed 

 during the course of other work on this portion 

 of the river that water temperatures in the 

 various pools, and localities within pools, follow 

 each other closely, at least when the average of 

 a period of several days is considered. Almost 

 everywhere the water is relatively shallow, is 

 subject to some current, and hence is well mixed. 

 There virtually is no vertical stratification and 

 comparatively little prolonged stagnation. TTie 

 effect of sudden changes in air temperature 

 upon the water surface temperature is largely 

 nullified by the current and wind and wave action . 

 The locality in question (Railroad Bridge) is near 

 to or actually a part of the main flow of the river 

 and hence produces a fairly representative 

 temperature sample. 



Table 21 records the weekly fishing suc- 

 cess averages for pool 10 and for the total (all 

 pools) . As shown by the table, there exists a 

 mathematical correlation between these two sets 

 of averages and the recorded water temperature 

 with a reasonably high degree of statistical sig- 

 nificance. Whether this correlation is actually 

 causal or is mostly coincidental may be open to 

 question since there are many other factors 

 which might enter the picture. If the relation- 

 ship of catch to water temperature, as shown 

 here, is a real one, the import is that within 

 the temperature range here recorded, the higher 



the temperature the better the fishing. The 

 findings of Dendy (1946), and of Eschmeyer, 

 Manges, and HasLbauer (1946) are to the same 

 effect; i.e., that within a certain range and under 

 a certain combination of circumstances, higher 

 water temperatures favor fishing. However, it 

 has been held by many that extremely warm 

 water may be the cause of the mid-summer slump 

 in fishing which sometimes occurs (perhaps in- 

 directly, by increasing the supply of available 

 food, through the hatching of insects) . 



Most of the localities, particularly such 

 ones as Lake Pepin, did not show a mathematical 

 correlation of fishing success with water temper- 

 ature. 



Changes in water levels. The moving, feeding, 

 and biting of most of the species of fish in the 

 river are influenced to a certain extent both by 

 the water level and by changes in level (rising 

 and falling) . This is the firm belief of most of 

 the sport fishermen and is based upon their ob- 

 servations . 



Information based on measurement, such 

 as the daily readings of water level gauges, is 

 difficult to apply to the changes of a large and 

 varied sport fishery. For instance, during the 

 summer of 1945 the hydrographs for the various 

 river stations (constructed from gauge readings) 

 showed a different pattern for the water levels 

 of the different pools, or even sections within 

 pools. Thus it is almost meaningless to attempt 

 to fix upon any mathematical correlations between 

 water levels (or level changes) and fishing suc- 

 cess. 



Throughout the entire area there was a 

 general trend from relatively high and fluctuat - 

 ing water levels (flood waters) at the beginning 

 of the summer fishing season, to lower and more 

 stabilized levels (normal pool levels) through 

 the latter half of July, August, and September . 

 As shown in figure 1, this change corresponds in 

 time to the increase in amount (also in fluctu- 

 ability) of the average fishing success. Whether 

 this relationship is a matter of cause and effect 

 or is purely coincidental is open to question. 



In the winters of 1944-45 and 1945-46 

 there occurred some rather drastic changes in 



39 



