AGE DETEKMIXATIOX FROM SCALES OF LAKE TROUT 



29 



records of culculatccl lengths for 97 lake trout 

 captured in large-mesh gill nets off Montague, 

 Mich., October 1, 1947 (Van Oosten 1950). 

 The scales of these fish were read by Dr. Frank W. 

 Jobes who did not record having observed the 

 0-mark. The calculated lengths from 82 of the 

 fish in the year classes 1939-43 yielded an average 

 length of 5.1 inches at the end of the first year of 

 life. This average is Ijctwecn (1.5 inches higher 

 and 1.0 inch lower than) the averages 3.6 and 6.1 

 inches obtained in tlie present study for the lengths 

 at formation of tiie 0-mark and the first annulus, 

 respectively, from 17 lake trout of the same year 

 classes from southern I..ake Michigan (off South 

 Haven in area 8) caught in the same year and 

 in nets of the same mesh size (table 23). These 

 differences suggest that on some scales Dr. 

 Jobes may have measured the first aniuilus to the 

 0-mark rather than to the first annulus. How- 

 ever, the calculated lengths " for the later years 

 of life of the lake trout from Montague and 

 South Haven were close enougli to indicate good 

 agreement on the assessment of age. 



From tliese data, it appears tlial without a 

 knowledge of the 0-mark, errors in measuring to 

 the first annulus of lake trout scales, due to mis- 

 interpretation of the central check, might be 

 numerous enough to bias seriously an estimate of 

 the first-year growth of I^ake Michigan lake trout, 

 but errors of age determination would be few. 



AGREEMENT BETWEEN FIRST AND SECOND 

 READINGS 



The two readings of lake trout scales, mentioned 

 previously, were made on the scales of all fish in 

 the collection. No samples were discarded, how- 

 ever difficult to read. The second series of read- 

 ings was begun several months after the first was 

 completed and, for each fish, a second scale was 

 read and measured, after comparison with the 

 other mounted scales in the sample. The two 

 readings agreed on age for 96.8 percent of the fish. 

 Errors of interpretation, not involving change in 

 age, reduced agreement to 91.4 percent of the 

 specimens. Because of experience gained during 

 the first reading, and standardization of proce- 

 dures, the second reading disclosed errors in the 

 earlier work as shown in table 8. Many of the 

 disagreements resulted from the omission of a 



'• Sums of the incrcmfnts of growtfi. Those for the lake trout from .Mon- 

 tapue. Mich., were obtained from the puhlislu'd data. 



measurement of the central or 0-mark, and the 

 mistaken location of annuli. However, there were 

 also disagreements on the number of annuli. 

 The number of annuli located during the second 

 reading varied from that recorded during the first 

 reading for 45 (3.2 percent) of the fisii as follows: 

 1 annulus more for 18 fish, 1 annulus less for 24 

 fish, 2 less for 2 fish, and 3 less for 1 fish. The 

 differences in percentage of such disagreements 

 among the data for tlu' tliree plantings were not 

 large. Disagreements in measurement, not re- 

 sulting in change of age, occurred for scales of 76 

 (5.4 percent) of the fish. 



T.\BLE 8. — Comparison of first and second readings by the 

 same person, of scales from the '^marked" lake trout 



Disagreements in readings due to omission of the 

 central check at the first reading were recorded, 

 but were not considered to be errors in reading 

 because the importance of measuring the 0-mark 

 was not fully understood at the beginning of the 

 first reading. Measurements of tlie central mark 

 had been taken commonly, liowever, when loca- 

 tion of the first annulus was math' easier In- defi- 

 nitely locating the central check. 



The scales of some lake trout present such prob- 

 lems of interpretation that readings made at differ- 

 ent times are likely to disagree. Much of this un- 

 certainty is dispelled by long familiarity with 

 scales from fish of known age. Most readers dis- 

 card the more didicult scales (usually about 5 per- 

 cent of the total) as unreadable. U this practice 

 liad been followed in the present study, some of 



