374 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



catch. This shift in age composition between the 

 two periods of study in Saginaw Bay is marked 

 by a similar change in the average age. The 

 average age was raised from 3.8 in 1929-30 to 

 4.3 in 1943-55. 



The Saginaw Bay yellow perch averaged older 

 than fish of other Great Lakes stocks for which 

 records of age have been published. The aver- 

 age age of Lake Erie perch was 3.2 years in 

 April. In southern Green Bay the average age 

 of 4.1 years in the spring was nearly equal to 

 the average of 4.3 for Saginaw Bay in 1943-55. 



The fall collection from Green Bay (average 

 age 2.7 years), northern Lake Michigan (3.6 

 years), and Lake Erie (2.2 years) were likewise 

 younger than those caught from Saginaw Bay 

 in October 1955 (3.7 years). It should be men- 

 tioned that in all three localities from which both 

 spring and fall collections were obtained the 

 spring fish were considerably the older. 



The effects of gear selection on the estimation 

 of average age cannot be judged precisely. 

 Catches of impounding nets (pound, trap, and 

 fyke nets) are probably comparable. The fall 

 samples from northern Lake Michigan, taken in 

 2%-inch-mesh gill nets, almost certainly were 

 biased by selection toward the older age groups. 



Relative Strength of Year Glasses 



Near the turn of the century, shortly after the 

 discovery and validation of the scale method of 

 determining the age of fish, investigators became 

 aware of wide variations from year to year in 

 the success of reproduction. As studies were 

 continued on individual stocks and expanded to 

 include new ones, it became increasingly appar- 

 ent that the fluctuations in the strength of the 

 year classes are major factors in the determina- 

 tion of the yield of the fisheries. It follows then 

 that an understanding of the extent and the fac- 

 tors in these fluctuations can contribute funda- 

 mentally to the development of a scientific system 

 of exploitation and management of fishery re- 

 sources. 



Despite the overriding importance of the sub- 

 ject and the considerable attention it has received, 

 the gaining of information on the magnitude of 

 fluctuations and the development of understand- 

 ing of their causes have been painfully slow. 

 The sampling requirements are rigorous, the 

 analytical procedures are difficult, and the fac- 

 tors that must be considered are numerous and 



complex. Seemingly we must approach under- 

 standing through a process of slow accretion in 

 which each new bit of evidence, though not con- 

 clusive in itself, must be welcomed. 



Fishery investigators in the Great Lakes mostly 

 have lacked the facilities for continuing studies 

 that are so essential for inquiries into the degree 

 and causes of fluctuations. Instead they have 

 had to content themselves with calling attention 

 to the occurrence of year classes of exceptional 

 strength or weakness. Observations of this type 

 have been made in Lake Erie for: yellow perch 

 ( Jobes 1933, 1952) ; walleyes, blue pike, and sau- 

 gers (Deason 1933) ; sheepshead (Van Oosten 

 1938) ; white bass (Van Oosten 1942) ; whitefish 

 (Van Oosten and Hile 1949) ; cisco (Scott 1951) ; 

 all of the species just listed (Van Oosten 1948). 

 In other Great Lakes waters, information on 

 year-class strength was recorded for: South Sagi- 

 naw Bay lake herring (Van Oosten 1929) ; Sagi- 

 naw Bay yellow perch (Hile and Jobes, 1941); 

 Saginaw Bay (Lake Huron) lake trout (Fry 

 1953) ; Green Bay walleyes, whitefish, and lake 

 herring (Hile 1950) ; Green Bay lake herring 

 (Smith 1956). In a 1954 paper, Hile attempted 

 the ranking of 12 consecutive year classes of the 

 Saginaw Bay walleye and commented on the 

 strength or weakness of others. 



Despite severe limitations and defects in the 

 data (small numbers of fish in some samples; 

 lack of collections in 1944 and 1952) it was be- 

 lieved that data on the age composition of yellow 

 perch in the present study have warranted an at- 

 tempt at a more precise estimate of the relative 

 strength of year classes than has been attempted 

 previously for Great Lakes stocks. It had been 

 hoped originally that these estimates might be 

 made through the application of sample data to 

 the statistical records on catch per unit of effort 

 in trap nets in the various years of capture, but 

 the scarcity (even lack) of legal-sized fish in the 

 samples forced abandonment of the idea. It was 

 then decided to use an adaptation of the proce- 

 dure employed by Hile (1941) for the estimation 

 of annual fluctuations in growth rate. It is based 

 on a series of comparisons in which the abun- 

 dance of each year class is estimated in terms 

 of the strength of the preceding one. From these 

 comparisons a sequence of positions is established 

 for each year class in the series. The procedure 

 can be illustrated by the comparison of year 

 classes 1943 and 1944* (table 10). The 1943 year 



