<br />generation to generation, one needs information at the same
<br />life stage (e.g., adult breeders) to encompass all factors that
<br />influence it. Nonetheless, if 33% of female razorback suckers
<br />contribute progeny each year, N of the 150,000 fish would very
<br />likely exceed the stated goal of 5% (50,000) of the ancestral
<br />1,000,000 fish. The annual yield of 10 progeny per female in
<br />our model maybe too conservative, but it was held at that level
<br />because of the uncertainties regarding N and concerns about
<br />production and harvest discussed below.
<br />We address these unknowns directly by promoting large,
<br />panmictic populations in the channel plus connectives, from
<br />which brood fish may be drawn. Doubling the number of
<br />hectares, females, or young per female in our model results
<br />in an order-of-magnitude change in estimated production.
<br />However, unlike these parameters (some of which, such as area
<br />dedicated for off-channel habitat, can be costly), altered sur-
<br />vivorship results in logarithmic adjustments in population size.
<br />An increase from 5% to 15% in first-year survival of repa-
<br />triates, for example, results in an estimated 40% increase in
<br />N in the channel plus connectives, from approximately
<br />54,000 to 5,000 and 150,000 to 210,000 at 5 and 20 years, re-
<br />spectively, in the examples given above. Increased survivor-
<br />ship, with a reduction in investment in the number of isolated
<br />habitats, brood fish, and other necessities, is thus the way to
<br />succeed.
<br />Completing the cycle. Step-like relationships exist between
<br />body size of repatriates at the time of release and their sur-
<br />vival in both rivers and reservoirs (figure 7). Thus, the young
<br />should be nurtured toward the largest possible body size be-
<br />fore they are transferred to the river. Given the major role of
<br />size-dependent predation in limiting reestablishment of Col-
<br />orado River fishes, transferring a few large (i.e., > 35 cm TL)
<br />individuals is far more productive than repatriating many
<br />smaller fish that would be devoured by predatory nonnative
<br />fishes.
<br />At moderate densities in lower-basin waters, young razor-
<br />back suckers can reach more than 30 cm and bonytail more
<br />than 25 cm TL in the fast year of life. By their second yeas',
<br />they commonly reach 45 cm and 30 or more cm TL, respec-
<br />tively. These growth rates allow for annual or biannual har-
<br />vesting. Growth rates of repatriates in Lake Mohave are sim-
<br />ilar, and substantial growth continues with increasing age
<br />(figure 8). As demonstrated by survival of repatriates (figures
<br />7 and 9), and in view of undetectable adult mortality in-
<br />ferred from long-term catch-per-unit efforts in Lake Mo-
<br />have between the 1970s and the onset of population collapse
<br />in the late 1980s (Pacey and Marsh 2003), razorback sucker
<br />of approximately 30 cm TL and longer are essentially immune
<br />to existing predators. First-year survival of approximately
<br />5% for repatriates in the wild increases to approximately
<br />80% in subsequent years (figure 7, top graph), correspond-
<br />ing roughly to increases in average TL from 30 to 35 cm at
<br />repatriation to 45 cm 2 to 3 years later (figure 9). These
<br />trends should apply as well for Colorado squawfish, although
<br />that species has a slower growth rate than razorback sucker
<br />Figure 7. Estimated survival of repatriated subadult
<br />razorback suckers in Lake Mohave (Arizona and Nevada)
<br />and San Juan River (New Mexico and Utah) that is based
<br />only on fish greater than 29 centimeters total length at
<br />time of stocking. The top graph ie based on raw data,
<br />with no assumptions applied. The bottom graph is based
<br />on the assumption that no individual is recaptured twice;
<br />as a result, all fish captured once are theoretically lost
<br />from the population.
<br />(Minckley et al. 1991, Osmundson et a1.1997). Chub, which
<br />more rarely exceed 35 to 40 cm TL as adults, may or may not
<br />be more vulnerable to such size-structured predation. What-
<br />ever the case, young should be nurtured toward harvest and
<br />repatriation at the largest possible body size.
<br />In practice, time of transfer from off-channel habitats is con-
<br />trolled by growth, which is partially a function of population
<br />size, which in turn is related to recruitment. Harvest and
<br />transfer should be balanced with production of appropriate-
<br />sized fish, but care must be exercised not to allow off-
<br />channel populations (brood fish plus progeny) to exceed
<br />carrying capacity, which would cause stunting. Potential dele-
<br />terious impacts on off-channel populations, such as parasitism,
<br />can be avoided by maintaining relatively low-density, highly
<br />productive stocks by appropriate harvest and translocation to
<br />230 BioSdence • March 2003 / Vol. 53 No. 3
|