LIFE HISTORY OF LAKE HERRING OF LAKE HURON 273 



because measured along the l)ody curvature. It is necessary to correct the labora- 

 tory measurements to those of fresh fish b}'^ making allowance for shrinkage in the 

 preservatives. 



Table 2. — Errors in the measurements of length of Bay City herring, due to use of the field method and 



due to shrinkage of the hody in formalin 



Milli- 

 meters 



.\verage length of 182 herring preserved in 4 per cent formalin lind measured Nov. 28, 1921, by field method | 238. 28 



.\verage length of the same 182 herring preserved in 4 per cent formalin and remeasured Nov. 28, 1921, by laboratorv 



methods-.. ' 232.98 



Difference due to method of meas'urement 



.Average length of 182 fre^h herring measured Oct. 26 and 27, 1921, by field method 



.Werage length of the same 182 herring preserved in 4 per cent formalin and remeasured Nov. 28, 1921, by field method 



DiUerence after shrinkage in formalin ._ _ 



Average length of 199 herring preserved Oct. 29, 1921, in 4 per cent formalin and measured Nov. 28, 1921, by field method.. 

 Average length of the same 199 herring preserved Oct. 29, 1921, in 4 per cent formalin and remeasured Dec. 2, 1921, by 



field method. 



Difference due to error in measurements made by field method , 



Coeflicient for correcting lengths of fish measured by the field method. _ 



Coefficient for correcting lengths of fish preserved in formalin 



5.30 

 242. 24 

 238.28 

 -3.96 

 238. 21 



239.54 

 d=1.33 

 .978 

 1.016 



To obtain a coefficient of correctiou for the length of fish measured along the 

 body curvature 182 preserved herring were measured on the same day by this method 

 and by the laboratory method. The difference between the averages, 238.28 for 

 the field method and 232.98 for the laboratory method (Table 2), was found to be 

 5. 30 millimeters. This mvolves an error of about + 0.022 millimeter for each 

 millimeter of the field measurement. The coefficient of curvature for correcting 



field measurements is therefore 17^9' or 0.978. To obtain a coefficient for shrinkage 



in preservation 182 fresh herring were measured by the field method and remeasured 

 by the same method after being a month in 4 per cent formalin. The difference of 

 tlie averages was 3.96 millimeters. (Table 2.) To learn whether this difference is 

 due to the errors involved in the method of measurement, 199 fish that had been in 

 formalin for about 30 days were measured and then remeasured after 4 days. The 

 difference between the two averages was found to be 1.33 millimeters. (Table 2.) 

 Thus, a fish of 242 millimeters apparently shrinks 3.96 ±1.33 millimeters through 

 the action of formalin, or 0.016 millimeter per milhmeter of body length. A shrink- 

 age coefficient of 1.016, therefore, was used for all preserved fish. 



Errors due to shrinkage or body curvature generally have been ignored in life- 

 history work. According to Ssemundsson (1913), Schmidt reports an average shrink- 

 age of 0.5 centimeter in the length of fish (plaice?) upon death. Williamson (1914) 

 observed no significant shrinkage in the length of the marine herring preserved in 

 ■formalm. Johanseu (1915) corrected his plaice lengths, as follows: 0.5 centimeter for 

 shrinkage at death and 1.0 centimeter when the specimen had begun to putrefy, had 

 been salted, or been left to dry. 



In the lake herring several other measurements besides total length were em- 

 ployed. H represents the length of the head as measured from the tip of the snout 

 to the most distant point on the margin of the bony suboperculum, excluding the 

 soft opercular membrane. Hi represents the length of the head as measured from 

 the tip of the snout to the most anterior point of the body proper on the projected 



