VETTER; NATURAL MORTALITY IN FISH STOCKS 



any given stock, abundant evidence exists to the 

 contrary. Natural mortality has been shown to 

 vary with age, density, disease, parasites, food 

 supply, predator abundance, water temperature, 

 fishing pressure, sex, and size. Evidence for rela- 

 tionships between these factors and M, and se- 

 lected references for each, are presented below. 



Changes in mortality rate with age, within sin- 

 gle groups of fish, have been demonstrated and 

 discussed more frequently than changes with any 

 other factor. References include, among others, 

 Baranov (1918, plaice), Sette (1943, Atlantic 

 mackerel), Ricker (1945, 1947, lake fish; 1969, 

 1975, various species), Beverton and Holt (1959, 

 many species of marine fish), Beverton (1963, en- 

 graulids and clupeids), Boiko (1964, sturgeon). 

 Gushing (1975, plaice), Blinov (1977, fish in gen- 

 eral), Bulgakova and Efimov (1982, Oregon hake 

 and sea perch), Sandland (1982, fish in general). 

 Smith (1985, clupeoids), Roff (1986, fish in gen- 

 eral). Evidence for changes with senescence for 

 fish in general has been discussed or documented 

 by, among others, Woodhead (1979) and Craig 

 (1984). 



Although specific patterns vary with species 

 (e.g., Woodhead 1979), in general M is extremely 

 high during egg and larval stages (e.g., 2 to 10% 

 per day in plaice and clupeoids (Cushing 1975; 

 Smith 1985)), falls precipitously during the juve- 

 nile period, becomes relatively stable during in- 

 termediate adult ages and increases again with 

 senescence. But even during these relatively sta- 

 ble mid-adult ages, changes in M with age can be 

 substantial, particularly in short-lived fish (e.g., 

 Ricker 1947, stunted versus "normal" whitefish). 



Changes in natural mortality rate with size 

 (rather than age) within single groups of fish 

 (usually stocks), have been discussed by Baranov 

 (1918, plaice), Ricker (1969, size-selective mortal- 

 ity in general). Ware (1975, larval fish), and 

 Peterson and Wroblewski (1984, many species). 

 Differences in natural mortality rate between 

 populations of the same species in different envi- 

 ronments, or even in different areas of a single 

 environment (e.g., a single lake) are documented 

 by Ricker (1947), Kennedy (1954), and Schupp 

 (1978). Year-to-year differences in natural mor- 

 tality rates of single stocks from a given area are 

 shown by Pope and Knights (1982, plaice) and by 

 Henderson et al. (1983, whitefish). Density- 

 dependent changes in M are discussed by Bever- 

 ton and Holt (1957), Cushing (1967), Tyler and 

 Gallucci (1980), Backiel and LeCren (1978), Jones 

 (1982), and others. Differences in M between 



sexes have been documented by Beverton and 

 Holt (1957, plaice), Ricker (1947, rock bass), and 

 others. Changes in natural mortality rate related 

 to the cost of reproduction have been discussed 

 by Jones and Johnston (1977), Roff (1984), and 

 others. 



Other factors that affect M either alone or in 

 combination with other factors include disease 

 and parasitism (reviewed by Lester 1984), starva- 

 tion (Hewitt et al. 1985; Theilacker 1986: larval 

 anchovy), physiological state (Smith 1985), and 

 fishing pressure (Ursin 1982; Munro 1982). Addi- 

 tional examples are cited by Beverton and Holt 

 (1957), Anderson and Ursin (1977), Sissenwine 

 (1984), and Hunter (1984). 



Most of the factors listed above (e.g., age, size, 

 sex) are indirect influences on M . The most im- 

 portant factor directly affecting natural mortality 

 rate is probably predation; this is implied by a 

 large body of literature describing changes in 

 prey community composition and abundance fol- 

 lowing changes in composition and abundance of 

 predators (e.g., Carpenter et al. 1985). 



Direct evidence that predators account for most 

 natural mortality in fish stocks is difficult to 

 gather (Section II). To quantify the fraction of M 

 due to predation, one must know, not only rela- 

 tive changes in abundance, but absolute popula- 

 tion density of all predators and prey together 

 with consumption rates and prey preferences of 

 all the predators. Although this is rarely possible, 

 at least two studies from freshwater systems do 

 present quantified estimates of predatory mortal- 

 ity in relation to available prey. Forney (1977) 

 quantified predation mortality in a relatively 

 simple, unmanipulated lake system where there 

 were few species of predator and prey. Combining 

 stomach-content estimates of prey consumed with 

 trawl-sample estimates of predator and prey 

 abundance, he concluded that 30 to 100% of yel- 

 low perch production was consumed by walleye, 

 their principal predator. In a manipulated sys- 

 tem. Stein et al. (1981) assessed predatory mor- 

 tality of young tiger muskellunge after they were 

 stocked in a small pond and lake. During the time 

 of the study, a single predator (largemouth bass) 

 accounted for 25 to 45% of losses to natural mor- 

 tality. 



In marine systems evidence for the relative im- 

 portance of predation can be gleaned from com- 

 paring total natural mortality with estimated 

 predatory mortality based on abundance of preda- 

 tors and feeding preferences. For example, multi- 

 species cohort analyses reported by Pope and 



35 



