DISCUSSION

Diaptomus kenai occurred across a relatively wide range of abiotic conditions, which is consistent with its status as the most common crustacean zooplankter in NOCA (Liss et al. 1995). Diaptomus kenai is apparently omnivorous (Olenick 1983; Butler et al. 1989), whereas Diaptomus arcticus is predaceous on rotifers and other crustacean zooplankton (Anderson 1967, 1970, 1972; Paul et al. 1995). Since Chaoborus rarely occurs in NOCA high lakes and cyclopoid copepods are seldom abundant, the two large diaptomid species are probably the top invertebrate predators in the pelagic region of lakes.

In NOCA, D. tyrrelli was found only in shallow lakes with relatively high levels of TKN and TP (TKN 0.05 mg l^-1, TP 0.007 mg l^-1). Byron et al. (1984) found that spatial variation of D. tyrrelli within Lake Tahoe was correlated with particulate nitrogen, but not with primary productivity, invertebrate predation, or algal biomass. Across its range, D. tyrrelli is not restricted to shallow lakes (Anderson 1974; Olenick 1983; Byron et al. 1984), as it appears to be in NOCA, although Olenick (1983) indicated that it was restricted to the upper 2-4 m in a relatively deep British Columbia lake.

Our results are consistent with the view that: 1) introduced trout can reduce or eliminate larger, more visible diaptomid taxa from lakes (Anderson 1974, 1980; Dodson 1979; Stoddard 1987; Starkweather 1990), 2) impacts on large diaptomids varies with fish density, with the greatest effects occurring at high fish densities (Brocksen et al. 1970; Stenson 1972; Langeland 1978; Dodson 1979; McQueen et al. 1986; Post and McQueen 1987) and, 3) impacts on diaptomids due to fish introductions are most likely to be observed in shallow lakes (Gliwicz and Prejs 1977; Donald et al. 1994).

Figure 4.4 is a hypothesis of inferred interrelationships of introduced trout and diaptomid copepods in higher-elevation lakes (>= 1200 m) in NOCA. Large diaptomid copepods were either absent or low in abundance in shallow lakes with high densities of reproducing trout. Large diaptomid abundance in lakes with low fish densities, most of which maintained non-reproducing trout, was significantly higher than in lakes with high fish densities, but was indistinguishable statistically from fishless lakes (with or without salamanders).

In deep lakes (maximum depth >10 m) with reproducing trout, large copepod densities were significantly higher than in shallow lakes. Perhaps densities of reproducing trout may not be as high in deep lakes as in shallow lakes, or large diaptomids may find refuge in deeper water during the day and so escape predation from a visually-oriented predator such as trout (Zaret and Suffern 1976; Stich and Lampert 1981; Gliwicz and Pijanowska 1988; Donald et al. 1994).

The small diaptomid, D. tyrrelli, occurred in five of seven lakes with high densities of trout where large copepods were absent or low in abundance. In these lakes the small diaptomid frequently reached high abundance. In contrast, D. tyrrelli rarely occurred in lakes with non- reproducing fish. In fishless lakes the small diaptomid was abundant only in lakes where large copepods were absent (Juanita and MR 2; Table 4.3).

The tendency for the small diaptomid to occur at high densities in lakes with high fish densities in part may be a consequence of interaction between large and small diaptomids. We observed a significant negative relationship between large diaptomid density and D. tyrrelli density for lakes with levels of TKN and TP 0.05 mg l^-1 and 0.007 mg l^-1, respectively. Although the mechanisms underlying this relationship are unknown, the negative correlation could imply interaction between large and small copepods. The replacement of large predatory copepods with smaller herbivorous copepods is consistent with Zaret's (1980) predation submodel II. In a lake in the Canadian Rockies, Anderson (1972) observed large increases in D. tyrrelli following elimination of D. arcticus by trout. Dodson (1974) and Paul et al. (1995) found significantly higher densities of a small herbivorous cyclopoid copepod in experimental enclosures without D. arcticus than in enclosures where D. arcticus was present.

Olenick (1983) indicated that D. kenai developed more rapidly in a British Columbia lake than did D. tyrrelli. We noted a similar pattern in NOCA; D. tyrrelli adults often did not appear until late August or September, whereas we commonly found adults of large diaptomids in late July and early August. Thus, in NOCA lakes where the species co-occur, D. tyrrelli nauplii could be vulnerable to competition or predation from late-stage copepodids or adults of large diaptomids. Naupliar stages of diaptomids may be particularly vulnerable to both predation and interspecific competition (Anderson 1970; Maly 1976; Olenick 1983; Paul et al. 1995).

Diaptomus tyrrelli occurred only in lakes with higher concentrations of TKN and TP. Leavitt et al. (1994) present evidence that introduced trout can increase phosphorous levels and algal biomass in high-elevation lakes. Thus high fish densities could favor D. tyrrelli in NOCA lakes both by reducing densities of large predatory copepods and by increasing nutrient levels in the pelagic zone. However, it is unlikely that high D. tyrrelli densities in lakes with fish are a consequence solely of increased nutrient cycling because the small diaptomid occurred at high densities in some fishless lakes.

The mechanism linking TKN and TP to the distribution of the small diaptomid is somewhat uncertain. TKN and TP are predictors of lake productivity (Lambou et al. 1983; Paloheimo and Fulthorpe 1987). Perhaps increased primary production resulted in greater food availability for the herbivorous small diaptomid.

Although fry were historically stocked at relatively high densities in Upper and Lower Panther Lakes, D. kenai was abundant in both lakes. Gliwicz and Pijanowska (1989) assert that reduction or elimination of zooplankton prey by vertebrate predators requires that predation intensity be "persistent." High predation intensity on large copepods may be relatively brief and episodic in lakes that are periodically stocked with high densities of fly if the interval between stocking is relatively long and fish mortality is high (Carpenter et al. 1985). This may be the case m the Panther lakes and it contrasts with lakes with reproducing trout where high fish densities and age and size diversity are more continually maintained. Furthermore, compared to higher elevation lakes, lower elevation lakes in NOCA tend to have greater diversity and abundance of nearshore benthic macroinvertebrates (Hoffman et al. 1996) and zooplankton (Liss et al. 1995), and moderately abundant populations of the salamander A. gracile. Thus, in contrast to higher elevation lakes, more alternative foods may be available for trout in lower elevation lakes and predation pressure on large copepods may be reduced.

ACKNOWLEDGMENTS

We are grateful to the former and present members of our Scientific Advisory Panel: Stanford Loeb, Stanley Dodson, Robert Hughes, William Neill, W. John O'Brien, James Petranka, William Platts, and H.B. Shaffer. They were extremely helpful in establishing direction for the research, reviewing annual reports and proposals, and suggesting improvements in research design and interpretation. Stanley Dodson and Nancy Butler provided constructive reviews of the manuscript. Notwithstanding, any shortcomings of the research are the sole responsibility of the authors. This research would not have been possible without the co operation and logistical support provided by personnel of North Cascades National Park Service Complex. The work was funded by the National Park Service and the USGS-Forest and Rangeland Ecosystem Science Center.