In this section, we summarize biological and environmental information that is relevant to the petition. Although we make particular efforts to address issues raised by the petitioners, we do not confine the evaluation to those issues alone. Information presented in this section forms the basis for conclusions regarding the species and threshold questions, which are addressed in the following section.
Steelhead may have the most variable life history of any anadromous Pacific salmonid species (Shapovalov and Taft 1954, Barnhart 1986). The complexity of steelhead life histories "reflect[s] extreme adaptability to a wide variety of environmental conditions" (Burgner et al. 1992, p. 9). The present northern and southern limits of spawning by steelhead are the Copper River basin, Alaska (63°N) and Malibu Creek, California (34°N), respectively (Burgner et al. 1992). Rivers utilized by steelhead exhibit considerable diversity with respect to geology, hydrology, stream flow, temperature regime, stream gradient, and biotic community structure.
Across their distribution, steelhead utilize river basins of all sizes. The Illinois River drains a watershed of approximately 2,565 square kilometers. Average discharge is 128 cubic meters per second, based on 1960-74 data (Forest Service 1977). The Illinois River enters the Rogue River at RKm 44 near the town of Agness, Oregon. Arising in.northern California, the Illinois River flows northwest through the Siskiyou Mountains of the Klamath Mountain Geological Province.
Klamath Mountain Geological Province
The Klamath Province includes a complex of mountain ranges in southwest Oregon and northwest California (approximately 40-43°N). Collectively, these are called the Klamath Mountains; they include the Trinity Alps, Salmon Mountains, Marble Mountains, and Siskiyou Mountains (Wallace 1983). Ecologically, the region is classified in the Marine Division of the Humid Temperate Domain (Bailey 1980); however, it exhibits influence from the warmer, drier Mediterranean Division (T. Atzet). This region includes diverse localized climates including cool, wet coastal areas and hot, dry interior valleys that receive less precipitation than any other location in the Pacific Northwest west of the Cascade Range (Franklin and Dyrness 1973). For example, average annual precipitation in the interior Rogue River valley ranges between 30 and 94 cm (Oregon Water Resources Committee 1955), while at Cave Junction in the middle Illinois Valley it is 152 cm, and Gold Beach at the mouth of the Rogue River receives 229 cm (Forest Service 1989).
The Siskiyou Mountains include northern extensions of geological formations typical of those found in the California Coast Ranges and the Sierra Nevada (Franklin and Dyrness 1973). The unusual geology and climate result in vegetation which "combines elements of the California, north coast, and eastern Oregon floras, with a large number of species indigenous only to the Klamath Mountains region" (Franklin and Dyrness 1973, p. 130).
Although it is true that the Illinois River is nearly in the southern fourth of the steelhead species' range (ONRC et al. 1992, p. 5), there are many steelhead-bearing streams south of the Illinois River, including the Klamath and Sacramento River basins. The Illinois River is located in a geologically and botanically unusual area; however, this area (the Klamath Mountain region) includes a number of rivers besides the Illinois River.
Across their distribution, steelhead are exposed to a wide range of water temperatures. In experiments, steelhead demonstrate an upper preferred temperature limit of 13°C, but they can survive water temperatures up to 24°C (Bell 1986). Temperature data collected in the Illinois River basin between 1989 and 1992 (Forest Service 1992b) indicate mid-day summer stream temperatures ranging between 11.1°C and 22.8°C in several Illinois River tributaries which are thought to have steelhead (Table 1)--generally above the preferred limit, but most well within the lethal limit. Dambacher (1991) recorded summer stream temperatures in the Steamboat Creek basin of the North Umpqua River, approximately 140 km northeast of the Illinois River. In 1987 and 1988, he found tributary temperatures from 10°C to over 20°C, and temperatures exceeded 25°C in the mainstem. Further, Everest (1973) reported that, prior to the construction of Lost Creek Dam in the 1970s, Rogue River summer steelhead often sought temporary thermal refuge in the relatively cooler water of the lower Illinois River. Therefore, present temperature readings for the Illinois River may not be indicative of long-term patterns.
There are, without doubt, locations within the Illinois River basin where summer water temperatures exceed the preferred range for steelhead. However, it is evident that relatively cooler areas also exist, with temperatures similar to those in other steelhead-bearing river basins located farther north, and that these areas have historically attracted steelhead from the Rogue River. It is possible that rather than (or in addition to) adapting to warm water, Illinois River steelhead seek out habitat that is within their temperature preference range.
Illinois River Falls
A notable feature of the Illinois River is a waterfall at RKm 64. Accounts differ as to whether steelhead succeeded in navigating the falls prior to any anthropogenic modification. However, a newspaper clipping dated December 1895 stated that "salmon" were going over the falls "in great numbers in search of spawning grounds in the headwaters of the stream." It is possible that these "salmon" were actually steelhead.
Hydrology of the Rogue River basin, including the Illinois River, shows a strong increase in discharge in December-January, which could make the falls naturally passable by steelhead. The ability of steelhead to navigate a waterfall is greatly affected by streamflow levels. A waterfall that is a complete barrier at one flow level may be easily passable at a higher or lower flow. Therefore, migration timing and basin hydrology are important factors for determining whether a waterfall is a migration barrier to a population of steelhead. Withler (1966) described a waterfall on the Coquihalla River in British Columbia that is a barrier to winter steelhead but not summer steelhead, and many steelhead bearing streams besides the Illinois River have waterfalls that may be seasonal barriers to anadromous fish passage.
The history of modification of the Illinois River Falls for fish passage is unclear. The earliest account is of miners' attempts to blast 3 feet of material from the top of the falls in 1910 (Palmisano 1992). A resident of the area stated that the falls were first modified by the State of Oregon in 1912 (Forest Service 1992a). Three to 4 feet of material were reportedly blasted from the top of the falls. The next account is of a fishway blasted on the south side of the falls in 1930 (Palmisano 1992). Rivers (1963) stated that a Forest Service biologist recommended and oversaw the blasting of 50 tons of rock from the falls in October 1935. Finally, a fish ladder was constructed in 1956-57 (Palmisano 1992).
Because of this uncertainty about the history of fish passage and alterations to the Illinois River Falls, we were unable to determine what effect, if any, the falls have had on reproductive isolation of Illinois River winter steelhead.
Steelhead exhibit variations in life history characteristics both across their distribution and within populations. Variable characteristics include age at smoltification, length of ocean residency, timing and duration of spawning migration, incidence of repeat spawning, and degree of residualization. Some of these variations have geographic trends. Fishery biologists group steelhead populations on the basis of life history, distribution, and genetic characteristics.
Steelhead may be grouped as coastal or inland populations (Behnke 1970). Allendorf (1975) reported genetic distinctions between steelhead populations of the Fraser and Columbia River basins (inland) and steelhead populations of basins west of the Cascade mountain range (coastal). Illinois River steelhead are included in the coastal group.
Steelhead utilize two main spawning migration strategies. Summer-run steelhead enter fresh water between May and October. Upon stream entry, their gonads are immature and secondary sex characteristics are lacking; summer steelhead spend up to 10 months in fresh water prior to spawning (Smith 1969). Winter-run steelhead migrate from November to April. They have well developed gonads at stream entry and spawn within 4 months of entering fresh water (Smith 1969). The Illinois River is generally considered to have only winter-run steelhead (but see below under Summer- and winter-run steelhead).
Freshwater Life History and Smoltification
Juvenile steelhead commonly spend 1 to 3 years in fresh water prior to smoltification and outmigration (Neave 1944, Behnke 1979), but up to 6 or more years in some cases (Peven 1990). Age at smoltification tends to increase among steelhead populations from south to north (Withler 1966, Burgner et al. 1992) (Table 2). Bali (1959) studied scales from 14 coastal Oregon winter steelhead populations. In each of these populations, over 70% of the steelhead smolted at age 2 or older. Scale data (ODFW 1992a) from 125 adult steelhead caught in the Illinois River from January 1982 to February 1990 indicate that 99% smolted at age 2 or older. This is not unusual in comparison with other coastal steelhead populations (Table 2); however, it is a somewhat higher age at smolting compared to Rogue River steelhead (ODFW 1990). We were unable to determine whether the relatively young age at smolting for Rogue River steelhead predated the effects of artificial propagation, which tends to accelerate smoltification.
Several studies have described the relationship of steelhead smolt size to ocean survival (e.g., Seelbach 1987, Ward and Slaney 1988). Burgner et al. (1992) stated that the average fork length for wild steelhead at smolting is about 160 mm for most populations. Bali (1959) backcalculated length at age for 620 winter steelhead from 14 coastal Oregon drainages; the smallest average smolt length was 162 mm for populations in the Wilson River and Sand Creek. No steelhead from the Rogue River basin were included in Bali's work. The only report of smolt size for Illinois River winter steelhead is from Rivers (1957), who stated that the average smolt was 147 mm. However, Rivers provided no data to support this statement, and it has been questioned by some ODFW biologists (T. Satterthwaite), whose data indicate that average smolt size for Rogue Basin winter steelhead ranges from 199 to 275 mm, depending on life history pattern (ODFW 1990). However, there is evidence in the literature of other steelhead populations with average smolt size below 160 mm, similar to that reported by Rivers for Illinois River winter steelhead (Maher and Larkin 1955, Chapman 1958; see Table 3).
Thus, although data are incomplete, it appears that steelhead from the Illinois River may on average smolt at a size near the low end of the range found in other coastal steelhead populations. Average age at smolting of Illinois River winter steelhead appears to be higher than for Rogue River steelhead but comparable to other coastal steelhead populations.
The steelhead life history pattern called the "half-pounder" (Snyder 1925) is limited in North America to the Rogue, Klamath, Mad, and Eel River drainages of southern Oregon and northern California (Barnhart 1986). Half-pounders return from the ocean to fresh water from July through September, after only 2 to 4 months of saltwater residence. They generally overwinter in fresh water before outmigrating again in the spring. There is some variability in criteria for defining half-pounders. Kesner and Barnhart (1972) described Klamath River half-pounders as being 250-349 mm. Everest (1973) used 406 mm as the upper limit of half-pounder body length on the Rogue River.
The half-pounder migration has been termed a "false spawning run" because few half-pounders are believed to be sexually mature. However, Everest (1973) found some spawning activity by male half-pounders. Most half-pounders found by Everest (1973) on Rogue River spawning grounds were 355-406 mm.
Half-pounders can migrate significant distances; for example, half-pounders of Klamath River origin have been found in the Rogue River (Everest 1973). It is apparently common for steelhead to make their half-pounder run into a nonnatal stream and then return to their natal stream to spawn as mature adults (Everest 1973, Satterthwaite 1988). A popular sport fishery has developed around the Klamath and Rogue half-pounder runs.
Half-pounders are generally associated with summer-run steelhead populations. However, this trait has also been identified in winter-run steelhead, albeit at a lower frequency. For example, Hopelain (1987) found a half-pounder frequency of 23.2% among lower Klamath River winter-run steelhead, as compared to a mean frequency of 95.2% among fall-run (summer) steelhead from six Klamath River tributaries. Scale analysis of Rogue River winter steelhead initially collected for Cole Rivers Hatchery brood stock indicated a half-pounder frequency of approximately 30% (M. Evenson).
Presumably, the half-pounder life history occurs either to avoid a deleterious condition in the ocean or to exploit a beneficial condition inland. However, since half-pounders were first described in the literature (Snyder 1925), little additional information has been published, and no convincing theories to explain half-pounders have been advanced. It is not known to what degree this trait is due to genetic as opposed to environmental factors. In initiating the winter-run steelhead brood stock at Cole Rivers Hatchery (on the Rogue River), scale patterns were used to select fish that lacked the half-pounder life history (M. Evenson). Recently, however, there is evidence of half-pounders among winter-run steelhead returning to the hatchery. Cramer et al. (1985, p. 112) stated that the
occurrence of the half-pounder life history has increased among winter steelhead released from Cole Rivers Hatchery since the time that growth rates of parr in the hatchery have been accelerated in order to produce age 1 smolts.
These findings suggest that the incidence of the half-pounder life history can be strongly influenced by environmental conditions.
Illinois River steelhead scale data from ODFW (1992a) indicate that of 163 steelhead angled between January 1982 and February 1990, 158 were mature adults and 5 (3%) were half-pounders. It is possible that the few half-pounders had roamed from their natal stream and were not of Illinois River origin. The ODFW data do not indicate whether any of the mature adults had scale patterns indicative of previous half-pounder runs.
Although half-pounders occur at a much lower frequency among Illinois River steelhead than Rogue River steelhead, the Illinois River is not unique among coastal steelhead streams in not having half-pounders. In fact, most steelhead populations coastwide do not have this life history trait. We were unable to determine whether other river basins besides.the Rogue River that have half-pounders (i.e., the Klamath, Mad and Eel Rivers) have tributaries like the Illinois River that lack the trait.
Saltwater Life History
Steelhead generally remain in salt water for 1 to 4 years before their first spawning migration. Withler (1966) in North America and Maksimov (1976) in Russia reported that, in general, residency period in the ocean increases across the species distribution from south to north. We found that 86% of Rogue River winter steelhead and 83% of Illinois River winter steelhead spent 2 years in the ocean prior to their first spawning migration. Data for other coastal Oregon steelhead populations showed considerable variation in saltwater age at first spawning (Table 4). However, these data were obtained in four different studies covering a wide range of years, so at least some of the interpopulational differences shown in this table may be attributable to these factors.
Information on the behavior of steelhead during their saltwater phase is limited. Pearcy et al. (1990) published observations based on 134 juvenile steelhead collected at sea from 1981 to 1985. They found that steelhead originating south of Cape Blanco, Oregon, rarely migrated north of that point. Burgner et al. (1992) interpreted data collected from 1955 to 1990 by research vessels of the United States, Canada, and Japan. Outmigrating smolts occurred in nearshore sampling in May, but by July they had moved offshore. The only nearshore area where steelhead in their first ocean year (age .0) remained by July was off northern California. These fish were interpreted to be half-pounders preparing to move into fresh water. Burgner et al. (1992, p 33) also noted that
a small concentration of age .1 juveniles appears in the spring off the coast of southern Oregon and northern California.... These fish are separated from the main mass of age .1 steelhead and are likely half-pounders returning to sea.
Burgner et al. (1992, p. 43) found that throughout the ocean phase,
steelhead from coastal Oregon and California may have more restricted westward migrations than do the more northern stocks.... Although these results may be an artifact of the lack of coded-wire tagged smolt releases from coastal Oregon and the relatively low number of releases from California, there may well be a true difference in ocean distribution of these stocks.
Although data are limited, authors studying ocean distribution of Pacific salmonids have described differences in ocean migration patterns among steelhead originating south of Cape Blanco, Oregon; this area includes the Illinois River. We are not aware of any ocean distribution data specifically for Illinois River steelhead.
Migration, or "run," timing varies both among and within steelhead populations. Several authors describe a heritable component to run timing which has facilitated selection for early maturing steelhead in some hatchery programs (Cramer et al. 1985). The heritability of run timing may allow steelhead populations to exploit or avoid environmental factors which are fairly consistent from year to year. An example is the phenomenon of lagoon formation in coastal streams of southern Oregon and northern California, such as occurs in the Pistol and Mattole Rivers. Sand berms close the mouths of these streams during summer low flow, making them inaccessible from approximately June to September each year. Migration timing must coincide with months when the berm is breached by streamflow and wave action.
Authors reporting the migration timing of steelhead generally use one or more of the following methods of data collection: reports from anglers, collection of fish entering fresh water (by seining or electrofishing), or observations of fish at a dam or fishway. Relatively few authors report the actual spawning time. This is due in part to the difficulty of observing.steelhead under the high streamflow conditions in which they generally spawn. Spawning surveys were attempted in the Rogue River basin, including the Illinois River, in the 1950s, but high water flow and difficult field conditions hampered the efforts (Fish and Wildlife Service 1956). We were not able to find any comprehensive spawning survey information for steelhead in the Illinois River.
Summer- and winter-run steelhead--Migration timing does not necessarily provide a good indicator of when steelhead spawn. Summer steelhead enter fresh water in a reproductively immature state, as early as May, and do not spawn for many months. "Winter steelhead" migrate when they are closer to reproductive maturity. Both summer and winter steelhead spawn in the winter to early spring. In drainages with sympatric populations of summer and winter steelhead, there may or may not be temporal or spatial separation of spawning. Everest (1973) described spawning for Rogue River summer steelhead as December-March and for winter steelhead as March-June. Large rivers such as the Klamath and Rogue may have migrating steelhead throughout the year (Kesner and Barnhart 1972). Often, there are peaks in migration which are used to describe different runs. The most commonly described are the summer and winter runs, with the names referring to the season in which the steelhead enter fresh water. Spring- and fall-run steelhead enter fresh water reproductively immature and, therefore, are grouped with summer-run steelhead in this document. Rivers (1957) described three runs of steelhead in the Rogue River basin: spring (early summer), fall (late summer), and winter. Within these runs, he described 11 geographic "races" (Table 5). These included the Illinois River winter steelhead and a very weak run of Illinois River fall (summer) steelhead. These summer steelhead may actually have been Rogue River summer steelhead that historically sought thermal refuge in the Illinois River (Everest 1973). The most current information on run timing for Rogue River basin steelhead comes from ODFW. "ODFW now views the Rogue River basin as having two runs of steelhead: a summer run that generally enters the Rogue River from April through October; and a winter run that generally enters from November through March" (Fustish et al. 1989, p. 4; see also Table 6).
At present, only winter steelhead are believed to utilize the Illinois River (M. Jennings). However, a 46-cm steelhead in Indigo Creek, 2 miles above the confluence with the Illinois River, was documented in August 1990 (Forest Service 1992b). The date would seem to indicate that this was a summer steelhead.
Spawn timing may be heritable at least in part, but it is also subject to modification by streamflow, water temperature, and other variables. Rivers (1963) stated that in years of average streamflow and water temperature, Rogue Basin winter steelhead demonstrated the following pattern in peak spawning activity: Rogue River, 15 March; Illinois River, 1 April; and Applegate River, mid- to late April. Notably, Everest (1973) found a 2-week shift in peak spawning between years for Rogue River summer steelhead.
Age and body size at first spawning--Although Shapovalov and Taft (1954) found some steelhead which spawned prior to outmigrating to the ocean, steelhead generally spawn after a period of saltwater residency. Age at first spawning is thus a function of both the age at smoltification and the duration of saltwater residency. Data indicate that age at first spawning for Illinois River steelhead is similar to that of other coastal winter steelhead populations in Oregon (Table 7). Based on interpretation of 232 scale samples provided by anglers, Rivers (1957) reported average lengths at first spawning migration for Rogue Basin winter steelhead: Illinois River, 60 cm; Rogue River, 44 cm; and Applegate River, 39 cm. For comparison, Shapovalov and Taft (1954) found the average length at first spawning for Waddell Creek (California) steelhead ranged from 39 to 78 cm, depending on life history pattern. Therefore, it is evident that there is considerable variability in the size a winter steelhead attains prior to its first spawning migration even within a given stream. Rivers (1957) stated that the largest steelhead caught by anglers in the Rogue River basin were Illinois River winter steelhead, which averaged 64 cm overall. Rivers (1957) stated that the average angler-caught Illinois River winter steelhead weighed 2.9 kg; he gave no weight for winter steelhead from the Rogue and Applegate Rivers.
Repeat spawning--Incidence of repeat spawning tends to decrease from south to north (Withler 1966), with much variation among populations. Up to five spawning migrations have been recorded for an individual (Bali 1959); however, more than two is unusual. The majority of repeat spawners are female, presumably due to the extended time and energy males spend on the spawning ground competing for and guarding females (Shapovalov and Taft 1954, Withler 1966, Barnhart 1986). Columbia River steelhead are essentially semelparous (Long and Griffin 1937), typically surviving only one spawning migration. The Oregon Department of Fish and Wildlife found a repeat spawning frequency for Rogue River winter steelhead of 14.5% (ODFW 1990). From ODFW data (ODFW 1992a), we calculated a 20% repeat spawning frequency for Illinois River winter steelhead.
Distribution of spawning in Illinois River--Distribution of juvenile fish, especially young of the year, may provide an indication of spawning distribution in streams for which spawning surveys are unavailable. Recent stream surveys (Forest Service 1992b) provide a.partial record of distribution of juvenile steelhead in the Illinois River basin. Juvenile steelhead have been found in tributaries to the lower Illinois River as well as above the falls. This suggests that spawning by steelhead is distributed throughout the basin. Biologists have evaluated Illinois River tributaries for habitat quality and salmonid production capability (Forest Service 1992b). They concluded that the Indigo and Silver Creek basins of the lower river are "the most important tributaries of the Illinois River drainage in terms of fish production, natural flow and cooler summer water temperature" (Forest Service 1992b). However, these surveys were primarily limited to tributaries below Illinois River Falls.
Rainbow trout are the resident, or nonanadromous, form of steelhead (alternatively, steelhead can be viewed as the anadromous form of rainbow trout). Both forms are part of the biological species Oncorhynchus mykiss, and the taxonomy of this species, as well as the relationship between rainbow trout and steelhead, has been studied for several decades (e.g., Kendall 1920, Snyder 1925, Behnke 1992). No set of morphological or genetic characteristics have been found that consistently distinguish the two forms. It seems likely that resident rainbow trout have arisen independently many times following colonization of new areas by the anadromous steelhead. Foote et al. (1989) reached a similar conclusion for the species O. nerka--that is, that the resident form (kokanee) has arisen from the anadromous form (sockeye salmon) several times.
Because the focus in ESA evaluations of Pacific salmon and steelhead is on identifying and conserving genetic resources that are important to the evolutionary legacy of the biological species (Waples 1991), it is important to consider the genetic relationship between resident and anadromous forms. With respect to this status review, if resident.rainbow trout share a common gene pool with steelhead, the two forms should be considered together as a unit.
Unfortunately, little is known about the genetic relationship between steelhead and rainbow trout in any systems in which they co-occur. Native rainbow trout exist in the Rogue River basin (Behnke 1992), but we could find no published or unpublished information about the relationship between resident and anadromous forms in this system. Because of this lack of information about rainbow trout, this status review will focus exclusively on steelhead. The relationship between the two forms should be reevaluated if substantial information becomes available about rainbow trout in southern Oregon coastal streams.
Annual hatchery production of steelhead on the west coast of North America has increased since 1960 from about 3 million juvenile steelhead to almost 30 million in 1987 (Light 1989). The majority of hatchery-produced steelhead (89%) are from the Pacific Northwest states of Idaho, Washington, and Oregon (Table 8), and this figure is dominated by steelhead from hatcheries concentrated in the Columbia River basin (Light 1989).
Releases in the Illinois River
There are no steelhead hatcheries in the Illinois River basin; however, ODFW does have records of steelhead from elsewhere being released there (Table 9). The primary source for steelhead released into the Illinois River has been the Cole Rivers Hatchery on the Rogue River.
Scales collected from adult steelhead angled from the Illinois River have been analyzed by ODFW to determine frequency of hatchery and naturally produced steelhead. Scale data from ODFW (1992a) indicate that 4% of 162 steelhead caught in the Illinois River from January 1982 to February 1990 were of hatchery origin. It is possible that these hatchery steelhead were strays from the Rogue River. Rogue River steelhead are known to stray temporarily into the lower Illinois River in large numbers (Everest 1973). The Illinois River is listed as complying with ODFW wild fish policy (Chilcote et al. 1992), which states that escapement of at least 300 breeding fish per spawning season must be maintained and that up to 50% of those fish can be of hatchery origin if genetically similar to the native stock, or 10% if they are genetically dissimilar. The Illinois and Chetco Rivers contain the only Oregon populations of winter steelhead considered by ODFW to be in compliance with this policy (Chilcote et al. 1992).
Protein electrophoresis--Numerous protein electrophoretic studies of population structure in coastal anadromous and resident O. mykiss have been published since the early 1970s. Allendorf (1975) first distinguished two major genetic groups of O. mykiss, inland and coastal, separated geographically by the Cascade crest. These two groups have large and consistent differences in allele frequency that apply to both anadromous and resident forms of O. mykiss; that is, rainbow trout east of the Cascades are genetically more similar to steelhead from east of the Cascades than they are to rainbow trout west of the Cascades. Subsequent studies have supported this finding (Utter and Allendorf 1977, Okazaki 1984, Schreck et al. 1986, Reisenbichler et al. 1992), and similar differences have been identified between O. mykiss from the interior and coastal regions of British Columbia (Huzyk and Tsuyuki 1974, Parkinson 1984).
Genetic differentiation based on timing of upstream migration in steelhead has also been investigated by allozyme analysis. Allendorf (1975) and Utter and Allendorf (1977) found that, like resident and anadromous forms, summer and winter steelhead of a particular coastal stream tended to resemble one another genetically more than they resembled populations of adjacent drainages with similar run timing. Later allozyme studies have supported these conclusions in a variety of geographical areas (Chilcote et al. 1980, Schreck et al. 1986, Reisenbichler and Phelps 1989), including studies on steelhead from the Rogue River (Reisenbichler et al. 1992). However, in each of these more recent studies, the summer-run stocks have had some extent of hatchery introgression and may not represent the indigenous population. Furthermore, in at least some cases, interpretation of the results is complicated by difficulties in determining run timing of the sampled fish.
There is also evidence to suggest that the overall population structure of steelhead in the Pacific Northwest has been affected by artificial propagation. Parkinson (1984) found substantial genetic differences among steelhead populations from adjacent drainages in British Columbia. Studies from Washington (Allendorf 1975, Reisenbichler and Phelps 1989) and Oregon (Hatch 1990, Reisenbichler et al. 1992) reported smaller differences between populations. Reisenbichler and Phelps (1989) and Reisenbichler et al. (1992) suggested that since both Washington and Oregon had far more extensive hatchery steelhead programs in the 1970s and early 1980s than did British Columbia, the relative homogeneity among populations in these states may be due to introgression of hatchery fish into naturally spawning populations.
Allozyme studies of Oregon steelhead, including some populations from the Rogue River basin, were reported by McIntyre and Schreck (1976), Hatch (1990), and Reisenbichler.et al. (1992). Hatch (1990) surveyed 13 protein-coding loci in steelhead from 12 hatcheries and 26 coastal rivers or tributaries in Oregon, including Lawson Creek, a tributary of the lower Illinois River. He also sampled two tributaries of the lower Rogue River (Lobster and Saunders Creeks). Hatch found evidence for a "north-south cline" in allelic frequencies in 5 of the 13 enzyme systems analyzed, but only in river systems larger than 350 km2. Hatch (1990; p. 17) also reported that "the area south of the Coos River was marked by sharp transition in four different enzymes..." and (p. 33) "the pattern of several alleles ending their detectable Oregon presence just north of Cape Blanco suggests that there is a less than average amount of straying between the populations north and south of this feature."
Data reported by Reisenbichler et al. (1992) are based on steelhead collected from 24 natural sites and 13 hatcheries in the Pacific Northwest in 1971-78 and analyzed for 10 gene loci. Twenty-four of the collections were from the Oregon coast, including eight localities in the mainstem Rogue River and its tributaries, but none were from the Illinois River. Collections were also made in northern California from the Mad River Hatchery (winter run) and the Trinity River (summer run).
Results from Reisenbichler et al. (1992) do not suggest a geographic cline of allele frequencies; instead, evidence was found for some genetic differentiation between clusters of populations. In their analysis, steelhead from north of the Umpqua River formed a separate cluster from steelhead in southern Oregon and northern California. Genetic differences between fish in separate drainages within clusters were not statistically significant and were similar in magnitude to those reported in coastal Washington (Allendorf 1975, Reisenbichler and Phelps 1989) and less than reported in British Columbia (Parkinson 1984). Significant.differences were detected between hatchery fish and naturally spawning populations, including Cole Rivers Hatchery and a number of natural stocks from the Rogue River.
Reisenbichler et al. (1992) found that samples from the Rogue River basin tended to group together (Cole Rivers Hatchery with the lower and upper Rogue River mainstem collections, and Galice Creek with Shasta Costa Creek and Big Windy Creek). There were two exceptions to this pattern: steelhead from Saunders Creek and Slate Creek were genetically more similar to samples from northern Oregon than they were to other Rogue River samples. The only natural sample from northern California (Trinity River) also grouped with the Rogue River fish. Steelhead from Mad River Hatchery grouped by themselves, separate from other natural and hatchery populations in California and Oregon.
Genetic differentiation between clusters of populations from the north and south Oregon coasts was also reported in a previous allozyme study (McIntyre and Schreck 1976). Collection sites were from 15 natural locations, including 1 at the mouth of the Illinois River, and 2 hatcheries. The authors examined seven polymorphic loci.
Chromosomal studies--Chromosome karyotypes in steelhead and rainbow trout have been extensively studied (see review in Thorgaard 1983). Chromosome numbers from 56 to 68 have been reported in O. mykiss, but Thorgaard (1983) found that a 58-chromosome karyotype was the most commonly observed karyotype in a survey of steelhead from Alaska to central California.
In contrast to results for studies of morphological and allozyme characters, Thorgaard's (1983) analysis did not reveal chromosomal differences between interior and coastal O. mykiss populations. All interior trout populations had predominately 58 chromosomes (Wilmot 1974; Miller 1972; Gold 1977; Thorgaard 1976, 1977, 1983) and most (but not all) coastal rainbow trout and steelhead populations also typically had 58-chromosome karyotypes (Wilmot 1974, Thorgaard 1983, Busack et al. 1980).
Thorgaard (1983) did detect a geographic pattern in chromosomal variability between some northern and southern sample sites. Although the 58-chromosome karyotype was predominant throughout the sampled range, there were two geographic regions characterized by steelhead with 59 or 60 chromosomes: the Puget Sound/Strait of Georgia region and the Rogue River/northern California region. However, the karyotypes of fish from these two regions were not the same. Northern fish with 59 or 60 chromosomes had a different number of subtelocentric and acrocentric chromosomes than did southern fish with 59 or 60 chromosomes (Thorgaard 1976). The groups may differ by a pericentric inversion or by the addition or deletion of heterochromatin. Even farther south, winter steelhead in the Mad and Gualala Rivers and resident trout in the San Luis Rey River commonly had 64 chromosomes.
Thorgaard (1983) also analyzed chromosomal variability in winter- and summer-run steelhead from the Quinault River (Washington) and the Rogue River. Chromosome number from these two river systems was different, but the chromosome number in summer and winter steelhead within each river system was similar. This is consistent with the previously discussed allozyme (e.g., Utter and Allendorf 1977) and morphological studies (Behnke 1992).
DNA analysis--Restriction endonuclease analysis of mitochondrial DNA (mtDNA) has been used to examine the structure of natural populations for over a decade (Avise et al. 1979, Brown et al. 1979). Although the mitochondrial genome in salmonids has also been studied (Wilson et al. 1985, Ferris and Berg 1987, Gyllensten and Wilson 1987), we have found no published information on DNA in steelhead from southern Oregon. The only such study we are aware of remains unpublished (N. Buroker). Buroker's study included 120 individuals from 23 major river systems in Alaska, British Columbia, Idaho, Washington, Oregon, and California. Steelhead from southern Oregon were found to be highly diverse in mtDNA clonal types. In the 120 fish analyzed, 18 mtDNA clones were observed. These clones were clustered into four lineages, all of which overlap in southern Oregon. The 12 fish examined from the Rogue River had 6 of the 18 mtDNA haplotypes observed throughout the study. In contrast, the Columbia River had relatively low levels of mtDNA haplotype diversity.
The above studies reported data for far fewer genetic markers (loci) than can be resolved using current techniques. Furthermore, in recent studies, only Hatch (1990) examined steelhead from the Illinois River, and then only from one location. To remedy these shortcomings in the dataset, NMFS biologists analyzed new samples of coastal steelhead from 13 natural and 2 hatchery populations, focusing on the Illinois and Rogue River drainages (Table 10; Fig. 2). The four Illinois River samples included two from the lower river (Lawson Creek and Indigo Creek), one from the middle river below Illinois River Falls (Briggs Creek), and one from the upper river (Grayback Creek).
Samples from natural populations were collected by electrofishing in September and October 1992 and consisted of juvenile fish 49 mm to 209 mm fork length. In general, it is difficult to determine the adult run timing of juvenile steelhead. Most of the samples are considered to represent winter-run fish because they came from streams that are believed to have only winter-run steelhead. In some cases, however, both summer and winter steelhead occur (e.g., in the Rogue and Klamath River drainages), and samples from these areas are considered to be of unknown run timing (Table 10). The Illinois River is often cited as having only winter-run steelhead, but there is some indication that summer-run fish may exist in the lower part of the river (Forest Service 1992b). Therefore, the two samples from the lower Illinois River (Lawson Creek and Indigo Creek) are also considered to be of uncertain run timing.
Protein electrophoresis followed procedures described by Aebersold et al. (1987), modified somewhat to take advantage of recent improvements developed by NMFS for the study of Snake River steelhead and by Washington Department of Fisheries for the study of Yakima Basin steelhead (Busack et al. 1991). Although 50 or more fish were collected from most populations, time and resources limited the genetic analysis to a maximum of 40 fish per sample. In addition to the 15 samples of coastal steelhead, a sample of 10 Snake River steelhead (from Camp Creek, a tributary of the Imnaha River in northeastern Oregon) was included as a control to help ensure that the locus and allele designations were consistent with those used in existing datasets.
Screening of 46 enzyme systems resulted in the collection of data for 69 presumptive gene loci that could be scored in all samples. Table 10 summarizes levels of genetic variability for 39 loci that were polymorphic (more than one allele present) in at least one sample. The percentage of loci that were polymorphic in a population ranged from 30.8% in Bandon Hatchery to 61.5% in Little Butte Creek and Winchuck River. Average heterozygosity over the 39 polymorphic loci ranged from 0.071 in Lobster Creek to 0.098 in Winchuck River (with a value of 0.103 found for the 1992 Snake River sample). These values are consistent with the relatively high levels of genetic variability reported for steelhead in previous studies.
To examine population genetic structure, we computed genetic distance values between each pair of populations based on the 39 polymorphic loci. The unweighted pair-group method with arithmetic averaging was used to cluster populations based on the matrix of genetic distance values. Genetic distances were computed using Nei's (1978) unbiased method, which includes a correction for sampling error in small samples. In addition to the 15 coastal populations and the 1992 sample of 10 Snake River steelhead, we included previously compiled data for a larger sample (99 fish) from the Snake River taken in 1990.
The dendrogram resulting from clustering the genetic distance values is shown in Figure 3. The large genetic difference between coastal and inland O. mykiss that has been reported by many previous authors is readily apparent: the distance between the two Snake River samples and the coastal populations (0.025) is an order of magnitude larger than the distances between most pairs of coastal populations. Within the coastal group, the three populations from north of Cape Blanco (Bandon Hatchery, Nehalem River, and Yaquina River) form a small cluster that differs from the more southerly populations at a genetic distance level of approximately 0.004. This is consistent with results reported by McIntyre and Schreck (1976), Hatch (1990), and Reisenbichler et al. (1992), who found evidence for some genetic differentiation between populations in northern and southern Oregon.
Another measure of genetic differentiation is Wright's (1978) FST, which represents the proportion of total genetic variance attributable to differences between populations. Considering only the coastal populations, FST was 0.038, similar to the value (0.034) found for Snake River spring/summer chinook salmon from the Grande Ronde, Imnaha, and Salmon River drainages (Waples et al., in press). Genetic distance values between the coastal steelhead populations were also similar in magnitude to those found between spring/summer chinook salmon populations in the Snake River (Waples et al. 1991 and Waples et al. in press). Snake River spring/summer chinook salmon are considered a single species under the ESA and are listed as threatened (NMFS 1992).
Apart from the north/south differences noted above, there is only weak evidence for further structuring of the coastal steelhead populations. Although statistically significant differences were found between each pair of samples when data for all gene loci were considered, the differences were not large in an absolute sense, and little geographic pattern is evident. The four samples from the Illinois River do not form a coherent genetic group. In fact, three of the four samples are genetically more similar to a sample from outside the Rogue River drainage than they are to other Illinois River samples (Table 11). The exception is Briggs Creek, for which Indigo Creek was the most similar population; however, the Indigo Creek sample was genetically closer to the Elk River sample than it was to Briggs Creek.