Discussion 

 Comparisons of Growth Measures 



Different procedures yielded different estimates 

 of growth rate. The two independent, direct mea- 

 sures based on recaptures — seasonal summation 

 and long-term recaptures — produced similar val- 

 ues (57 mm/yr and 62 mm/yr, respectively). The 

 indirect measures — length-age regression and 

 mean-length-at-age analysis — both projected an- 

 nual growth rates of 44 mm/yr. All estimates are 

 complicated by extreme variability in growth, 

 with overlap in lengths among four to six year 

 classes common (Smith and Saunders 1955; Ogden 

 1970; Gray and Andrews 1971; Hurley 1972). 

 Growth rate estimates based on recapture data 

 were apparently higher than those derived from 

 length-at-age analyses, but confidence intervals 

 overlapped among all estimates. However, we feel 

 that the direct measures are more accurate. First, 

 the sample size for the length-age analyses was 

 more than three times larger than for the seasonal 

 summation analysis, but the confidence intervals 

 were very similar (17.4 mm and 17.8 mm, respec- 

 tively), suggesting less variability in the recap- 

 ture data. Second, growth rates derived from 

 recaptured animals are based on actual growth 

 between captures; variability in calculated growth 

 rates should therefore reflect real variability in 

 growth among animals. In length-age analyses, 

 age classes are commonly resolved at no finer than 

 an annual level. Consequently, growth subse- 

 quent to day 1 of each year increases the variance 

 around the estimate rather than increasing the 

 accuracy of the estimate. Finally, the accuracy of 

 age determinations from otoliths in some eel 

 populations is questionable (Moriarty and Stein- 

 metz 1979; Deelder 1981; Casselman 1982), placing 

 length-age analyses in doubt unless annulus for- 

 mation can be verified. 



Limited growth data from other mark-recapture 

 studies of American eels are available. Hurley 

 ( 1972 ) tagged 1,418 American eels in Lake Ontario, 

 Canada, and reported recapture intervals for 13 

 large individuals (730-874 mm), which increased 

 an average of 34 mm/yr. At two Louisiana fresh- 

 water locales, Gunning and Shoop (1962) tagged 

 43 American eels; only four recaptures provided 

 usable data, indicating an average growth of 140 

 mm/yr (growth range = 46-325 mm, initial 

 lengths = 255-915 mm). R. L. Haedrich 2 tagged 



148 American eels in a Massachusetts estuary. 

 Four individuals (initial lengths = 500-700 mm) 

 had an average annual growth rate of 6% (range = 

 4.1-8.4%). An inverse latitudinal trend in growth 

 is suggested (see also Harrell and Loyacano 1980), 

 but direct comparison is complicated by different 

 initial lengths, small sample sizes, and high 

 variability in growth. 



Length-related differences in growth have also 

 been found for other populations. A shift from 

 allometric to symmetric growth occurred at 800 

 mm for American eels in Lake Ontario (Hurley 

 and Christie 1982). Those authors, as well as 

 Smith and Saunders (1955), related such a growth 

 change to physiological preparation for matura- 

 tion and migration. Gray and Andrews (1971) 

 found that American eels in New Brunswick, 

 Canada, estuaries grew slowly after age XI. Helf- 

 man et al. (in press) suggested that maturation of 

 Fridaycap Creek eels occurred at around age IV 

 (mean length = 387 mm). An apparent decrease in 

 growth rates of Fridaycap Creek animals longer 

 than 400 mm (Fig. 1) supports their interpretation. 



Causes of Seasonal Differences 



Seasonal and annual differences in growth rate 

 can be linked to fishing success as affected by 

 climate. Eel fishing in Georgia estuaries is typi- 

 cally poor at water temperatures below 10° C and 

 above 24° C. In 1980-81, estuarine water tem- 

 perature fell below 10° C during December 1980, 

 but average 1981-82 monthly temperatures were 

 higher and did not reach the 10° minimum until 

 January 1982. In addition, rainfall in 1981 was 45 

 cm below average, and mean water temperatures 

 were 2°C higher during June through September 

 than in 1982 (R. Arnsdorff 3 and T. E. Targett 4 ). 

 The winter slow growth period may therefore 

 result from colder water temperatures and re- 

 duced feeding. Faster growth in 1981 than in 1982 

 may have resulted from elevated temperatures 

 during much of the fast growth period of 1981. 

 High water temperatures — leading to reduced 

 feeding, interrupted growth, and poor fishing 

 — occurs in European eels, Anguilla anguilla 

 (Deelder 1981). Interrupted summer growth may 



R. L. Haedrich. Department of Biology, Memorial University 



of Newfoundland, St. John's, Newfoundland, Canada A1B 3X9, 

 pers. commun. April 1983. 



3 R. Arnsdorff, Georgia Department of Natural Resources, 

 Environmental Protection Division, 270 Washington St. S.W, 

 Atlanta, GA 30334, pers. commun. October 1982. 



4 T. E. Targett, Skidaway Institute of Oceanography, P.O. Box 

 13687, Savannah, GA 31406, pers. commun. October 1982. 



521 



