9640 - Laserfiche WebLink

Home Browse Search

tideq past. Since substantial genetic variation remains, it is prudent to perpetuate it, and with a suffi- ciently large N (the larger the better) generation of novel variation will continue. Management should be under the most natural condition possible, emphasizing achievement and main- tenance of species' carrying capacities in diverse habitats. The result will be increased opportu- nities for emergence of novel variation, thus maximizing adaptive potential. Demographics. Because of their great repro- ductive potential, conservation of large- bodied, long-lived fishes differs fundamentally from conservation of large-bodied, long-lived terrestrial vertebrates. Unlike most dry land ver- tebrates, almost all fishes produce great numbers of gametes-104 to 106 ova per female are not Table 3. Estimates of mitochondrial DNA (mtDNA) variation in three Colorado River fishes, with maximum likelihood estimates of long-term effective female population size. N., if the population is assumed constant over evolutionary time. Also given is the estimate of Ne if the population is allowed to grow or contract over evolutionary time and the direction of that change (used with permission from Garrigan et al. 2002). Species Humpback Razorback chub, Borrytail, sucker, Data and estimates Little Colorado Lake Mohave Lake Mohave Data cytb mtDNA gene ND2 ND2 Number of nucleotides 790 763 311 Sample size 18 16 49 Number of haplotypes 5 3 10 Estimates N. (constant size) 97,500 89,500 669,000 Ne (growth) 149,000 61,900 940,300 Growth Stable Declining Expanding unusual. Survival from egg to adult is, however, highly variable and typically low. Natural recruitment can be 0.01% or less. For razorback suckers living in communities with predatory nonnative fishes, recruitment failed for approximately 40 to 50 years because of the near-total loss of juveniles. Thus, a declining fish fauna commonly remains individual- rich while becoming species-poor. This paradox, absent in most terrestrial vertebrates, presents a great advantage for the manager who can devise ways to exploit the high reproduc- tive rate while still maintaining genetic variability. Each female razorback sucker bears an average of 1700 ova per cm stan- dard length (SL), and average SL in Lake Mohave was ap- proximately 50 cm in 1983-1984 (i.e., 85,000 ova per female; Minckley et al. 1991). A manager can consider this at two extremes, one from a simple view of production and the other incorporating genetic concerns. If they all survived, the progeny of a single large female would more than replace the entire population of approximately 73,500 adults estimated for Lake Mohave in the period 1980-1993. Theywould, how- ever, all be siblings (or half-siblings if more than a single male was involved), reducing genetic variation in a single event from high (Dowling et al. 1996b) to dangerously low. Alternatively, 0.005% survival (two young per female) of off- spring produced by approximately 36,750 females (half the 1980-1993 N of adults in Lake Mohave) would also replace the whole population, at the same time preserving the ex= isting genetic variation. Natural conditions obviously fall between these extremes. Long-lived species frequently reproduce in alternate years or even less frequently, so numbers contributing each year to a next generation may be only a fraction of N (see above). However, even if relatively few fish spawn each year, asyn- chronous spawning across years means that, over a repro- ductive life of more than 35 years, most if not all adults may well contribute to future generations. To most closely mimic such natural reproductive processes, and thereby retain existing and promote novel genetic variation, management must aim for the largest possible panmictic population, perpetuated by the highest possible M How can such a large, genetically diverse population be achieved for endangered species inhabiting a highly modified river? Hatcheries are a poor choice because spatial constraints may excessively restrict genetic diversity within managed populations (e.g., Hedrick et al. 2000). Perpetuating these fishes requires exploiting both the reproductive potentials of the fishes themselves and the continuing strict regulation of the Colorado River. Below we outline a plan that is grounded in biology, hydrology, and engineering and involves relatively modest funding. Off-channel habitats for conservation Part of our proposal's rationale is to avoid competition with sport fishermen. Traditional off-channel angling areas such as backwaters and ponds behind levees cannot realistically be expropriated, so new habitats-exclusively for native species- must be provided. Costs are also a concern, so these habitats should be secure, simple, and low maintenance, and they should exclude nonnative predators while providing ade- quate physicochemical and other conditions for life history requirements of natives. We envision a series of excavated habi- tats (figure 5) resembling the pristine lower Colorado River floodplain--isolated oxbow lakes and backwaters--as primary components of dedicated off-channel complexes. A success- ful prototype has been developed on the USFWS Havasu National Wildlife Refuge at Beal Lake, Arizona. Nonnative fishes can be excluded by passing water through size-graded gravel at inlets and outlets of excavated habitats. Proximity to the river will allow easy construction and main- tenance access, shorten travel distances, and allow exploita- tion of gravity flow and high water tables. An elevation dif- ference (head) between inflow and outflow is critical to promote current, to mirdmize water quality problems by ex- changing water, and to avoid the cost and unreliability of pumps. A river bend would be ideal, shortened by dredging March 2003 / Vol. 53 No. 3 • BioScience 227