Aspects of juvenile rainbow trout (Oncorhynchus mykiss) diet in relation to food supply during summer in the lower Tongariro River, New Zealand

Transcription

1 New Zealand Journal of Marine and Freshwater Research ISSN: (Print) (Online) Journal homepage: Aspects of juvenile rainbow trout (Oncorhynchus mykiss) diet in relation to food supply during summer in the lower Tongariro River, New Zealand M. Dedual & K. J. Collier To cite this article: M. Dedual & K. J. Collier (995) Aspects of juvenile rainbow trout (Oncorhynchus mykiss) diet in relation to food supply during summer in the lower Tongariro River, New Zealand, New Zealand Journal of Marine and Freshwater Research, 29:3, 38-39, DOI: 0.080/ To link to this article: Published online: 30 Mar 200. Submit your article to this journal Article views: 52 View related articles Citing articles: 3 View citing articles Full Terms & Conditions of access and use can be found at

2 New Zealand Journal of Marine and Freshwater Research, 995, Vol. 29: /95/ $2.50/0 The Royal Society of New Zealand Aspects of juvenile rainbow trout (Oncorhynchus mykiss) diet in relation to food supply during summer in the lower Tongariro River, New Zealand M. DEDUAL Department of Conservation Private Bag Turangi, New Zealand K. J. COLLIER Department of Conservation P.O. Box Wellington, New Zealand Present address: National Institute of Water & Atmospheric Research Ltd, P.O. Box -5, Hamilton, New Zealand Abstract Invertebrates were collected from the benthos, drift, and stomachs of juvenile rainbow trout (Oncorhynchus mykiss) in different flow environments and at different times of day in the lower Tongariro River in December 992, to investigate spatial and diel patterns in prey abundance and diet. The benthic and drift communities were dominated numerically by Diptera (both 7%), Oligochaeta (22-23%), and Trichoptera (5% and 2% in the benthos and drift, respectively). Terrestrial invertebrates comprised 3% of the drift. The most common prey in the stomachs of juvenile rainbow trout (44-30 mm fork length) were Diptera (74%), Trichoptera (9%), Ephemeroptera (6%), and some terrestrial organisms. Relative abundances of different invertebrate taxa in the benthos, drift, and stomachs of juvenile rainbow trout were all significantly intercorrelated. Juvenile trout fed selectively on Trichoptera (particularly emerging adults), the ephemeropteran Deleatidium spp., and some Diptera (mostly Maoridiamesa and Aphrophila neozelandica), and avoided Oligochaeta. The M94043 Received 7 August 994; accepted 5 March 995 stomach fullness index was similar during three periods between dawn and dusk, indicating that feeding activity was continuous. The proportion of Diptera in the diet of small fish was higher than in larger fish and the reverse was observed for the proportion of Trichoptera. Densities of most benthic invertebrate taxa favoured by juvenile trout were highest in medium or fast flowing habitats, suggesting that maintenance of such conditions is important for food production. Keywords rainbow trout; Oncorhynchus mykiss; benthic invertebrates; drift; feeding; prey selection; Tongariro River INTRODUCTION The Tongariro River in the central North Island of New Zealand constitutes the principal inflow to Lake Taupo and is an integral part of a nationally important hydro-electric power scheme. The river also supports a world-famous rainbow trout fishery (Stephens 989; Cryer 99). It is believed that the production of adult trout in Lake Taupo is mainly limited by the rearing capacity of the natal streams rather than by the size of the spawning run (Department of Conservation unpubl. data). Analyses of scales taken from rainbow trout in spawning runs in the Tongariro River show that the fish are using the river as rearing habitat, at least for one summer before migrating to Lake Taupo (Stephens 989; Pitkethley 990). The rearing habitat must provide both space and food, although the primary regulator may change seasonally. The space-food relationship has been found to be most important in spring, summer, and early autumn, whereas in winter "desirable" space alone may govern density (Hartman 963; Chapman 966; Bjornn 97). Recent work by Angradi & Griffith (989) suggests that wild rainbow trout feed mostly during the day in summer. In the Tongariro River, densities of juvenile rainbow trout and benthic invertebrates are greatest

3 382 New Zealand Journal of Marine and Freshwater Research, 995, Vol. 29 in December (Stephens 989; Quinn & Vickers 992). Stephens (989) suggested that, as food was very abundant at this time of the year, space could limit juvenile rainbow trout densities. To gain further insight into spatial and diel patterns of prey abundance and the diet of juvenile rainbow trout, we examined the contents of juvenile trout stomachs collected over three periods between dawn and dusk and in different flow environments; investigated the composition of the benthos in different flow environments and the drift at different times of the day; and determined the food preferences of juvenile rainbow trout. METHODS Study site The study was carried out on 7 December 992 in "Judges Pool" in the lower part of Tongariro River (NZMS260 T9: ). During the sampling, the flow of the river was 29.2 m 3 s" and the water temperature ranged from 2 to 3 C. The flow regime and physical characteristics of this section of the Tongariro River have been extensively described by Stephens (989) and Quinn & Vickers (992). Collection of samples Benthic invertebrates were taken with a Surber sampler (0. m 2, 0.5 mm mesh) by stirring the bottom to a depth of about 0 cm and brushing stones with a stiff nylon brush. Five samples were taken from "slow" ( m s^ ), "medium" ( m s" ) and "fast" ( m s" ) flowing habitats at depths ranging from 0.3 to 0.36 m. Water velocity above each Surber quadrat was measured at 0.4 times the depth from the bottom with a Scientific Instrument Model 205 current meter fitted with a Stewart Stream Gauging Counter. Drifting organisms were collected over three periods: dawn (0430 to 0700h NZST), noon (0930 to 200 h NZST), and dusk (700 to 930 h NZST) below the fast-water habitat where benthic samples were taken. The sampler was fitted with a 0.5 mm mesh net, had an aperture of 30 mm x 55 mm, and collected drift in the top 0.25 m of the water column. Water velocity at the aperture of the sampler varied between 0.62 and 0.69 m s ~ '. All macro-invertebrates were preserved in 0% formalin. In the laboratory, the samples were passed through mm and 0.5 mm mesh nested sieves. Material retained in the mm mesh was picked through by eye on a white tray. Material retained by the 0.5 mm sieve was sub-sampled by half and sorted under a binocular microscope (lox magnification). Samples of 7-20 juvenile rainbow trout were taken close to the benthic and drift sampling points with an electric fishing machine in "slow" water at dawn (0630 h), and in "fast" water at dawn (0645 h), noon (200 h), and dusk (830 h). This sampling strategy enabled us to compare the composition of invertebrates in the drift or benthos with those in the stomachs of juvenile trout collected nearby. The fork lengths (FL in mm) and weights (W in g) of the fish were measured. The relationship between weight and fork length was described by the regression equation: W = a FL b where a and b are constants. The fish were killed with an overdose of benzocaine, their body cavities opened, and whole fish were preserved in 0% formalin. Later, the stomachs were removed and their contents stored in 0% formalin until identification. The dry weight of the stomach contents was measured to the nearest mg after 4 h desiccation at 68 C. Data analysis Differences between the mean relative abundance of each invertebrate order present in the benthos at different flow velocities were tested by one-way analysis of variance and a posteriori means comparison by Tukey's test. Differences in relative abundance of each invertebrate order present in the stomachs of juvenile trout caught in "fast" water at dawn, noon, and dusk were tested by oneway analysis of variance. Relative abundance of each invertebrate order present in trout stomachs at dawn in "fast" and "slow" water were compared with a Mest. The same procedure was used to compare the mean fork length of the fish samples. The relationship between the mean relative abundance of different invertebrate taxa in the benthos, drift, and stomachs of juvenile trout were assessed by Spearman rank correlation (r s ) analysis with a correction for tied data after Zar (984). Differences between the r s values were established with a Z-test (Zar 984).

4 Dedual & Collier Food of juvenile rainbow trout 383 The relationship between juvenile trout fork length and the percentage of major insect orders found in their stomachs was described by a regression analysis. As the amount of food in the stomach of a trout is likely to increase with size of trout, we used the gut fullness index (GFI) of Wurtsbaugh&Li(985): GFI = 0 5 W FL" where W = dry weight of stomach contents (mg), FL = fish fork length (mm), and b - the exponent of the fork length-weight regression of fish. Differences between GFI values for juvenile trout caught in "fast" water at dawn, midday, and dusk were tested for significance by Kruskall-Wallis analysis of variance. Differences in GFI at dawn between fish from fast and slow water were tested for significance with a Mann-Whitney test with tied rank (Zar 984). Prey selection of the more commonly eaten prey was examined by Pearre's (982) prey selection index (V): O t a s a. 80 -, v = -a.b H The significance of V was tested with a X 2 evaluation (X 2 = (adb e -a e b(j) 2 «/abde) where ad and a e are the numbers of a given prey taxon in the diet and environment, respectively; b d and b e are the numbers of all other prey taxa in the diet and environment respectively; and a = ad + a e, b = bd + b e, d = ad + bd, e = a e + b e, and n - a d + a e + bd + b e. Values of V can range from - to +, representing complete avoidance and high selection, respectively, whereas 0 implies no selection. To test if the V values for the most preferred prey were changing with time we used a Kruskall-Wallis analysis of variance. To detect any change in V values at dawn between fast and slow water we performed a Mann-Whitney test with tied rank when required (Zar 984). For all statistical tests, the significance level was set at P < RESULTS Invertebrate communities The numbers of the different invertebrate taxa in the benthos and drift at Judges Pool on the December sampling date are presented in Appendix. The benthos was dominated by Diptera (7%; all flow microhabitats combined), followed by Oligochaeta (22%), Trichoptera (5%), and Fig. Percentage composition of invertebrates collected in the benthos at different flow conditions (see text: Collection of samples) at Judge's Pool, Tongariro River, on 7 December 992. * significantly different at P < E, Ephemeroptera; P, Plecoptera; T, Trichoptera; C, Coleoptera; D, Diptera; O, Other. Ephemeroptera (%). Coleoptera, Plecoptera, and Megaloptera combined accounted for less than % of the benthic community. The benthos contained significantly more Oligochaeta in "slow" water than in "medium" or "fast" water (P = 0.009), and more Tricoptera (mostly Aoteapsyche spp.) in "fast" water (P = 0.008): see Fig.. A very similar pattern of relative abundance appeared in the drift, for all diel periods combined, where Diptera accounted for 7%, Oligochaeta 23%, Tricoptera 2%, and Ephemeroptera < % of total invertebrates collected. Terrestrial invertebrates contributed up to 3% of the drift in different samples (Fig. 2). Drift density peaked during the dawn period at 7.5 individuals irr 3, dropped during the noon period to 4.8, and increased again during the dusk period to 7.2 individuals m~ 3. Juvenile trout The fork length (FL) of the fish collected ranged from 44 to 30 mm, but 96% were < 5 mm. The

5 384 New Zealand Journal of Marine and Freshwater Research, 995, Vol. 29 all the fish caught were feeding. The mean GFI for the juvenile rainbow trout caught during the dawn period in "slow" water was 4.62 (Table ). In "fast" water the mean GFI peaked at dawn (3.50), dropped at noon (2.36), and rose again at dusk (3.6). However, none of these variations were significant (H = 3.34; X 2 crit = 7.85; < P < 0.0 and U = 79; U crit. = 225; 0.0 < P < 0.02 for Kruskall-Wallis and Mann- Whitney tests, respectively: Table ). Fig. 2 Percentage composition of invertebrates collected in the drift at dawn ( h NZST), noon ( h), and dusk ( h) (see text: Collection of samples) at Judge's Pool, Tongariro River, on 7 December 992. E, Ephemeroptera; P, Plecoptera; T, Trichoptera; D, Diptera; O, Other; Tr, Terrestrial. mean length of the fish caught at dawn in fast water (5 +.5 mm FL) was significantly smaller than of those caught at other times or in different flow habitats (Table ). The relationship between weight (W) and fork length (FL) for all fish combined was described by the regression equation W = 79x 0-7 FL 3l38 (r 2 = 0.98, n = 74). We did not find any fish with empty stomachs over the whole period of sampling, indicating that Trout stomach contents The composition of stomach contents of juvenile trout collected during the day is presented in Appendix 2. Pupal and larval Diptera (predominantly Chironomidae) were the most abundant prey consumed by juvenile rainbow trout in all samples. Trichoptera (mostly Aoteapsyche and adult caddisflies) was the second most numerous prey taxon, and was significantly more numerous in the stomachs of trout collected at dawn in "slow" water than in "fast" water (P = ) (Fig. 3). Trout caught at dawn in "slow" water had significantly more Ephemeroptera in their stomachs than those caught during the same period in "fast" water (P = 0.003) (Fig. 3). Terrestrial invertebrates comprised.4 0% of trout stomach contents. Relative abundances of this group did not change significantly either between flow microhabitat (Fig. 3) or between sampling time, although the proportion was considerably higher at dusk (Fig. 4). Plecoptera were found only in the stomachs of six fish caught at midday and three caught at dusk. Oligochaeta were not recorded from any fish. Rank correlation coefficients between relative abundances of the different invertebrate taxa present in the benthos and trout stomachs (r s = 0.70), and in the drift and trout stomachs (r s = 0.89) were all significant (P < 0.00). A significantly higher Table Mean fork-length (FL) and gut fullness indices (GFI) of the different fish samples (see text: Collection of samples) with their related standard errors and samples sizes («). * significantly different from any other samples (P < 0.05, Tukey-test). Habitat Diel period FL (mm) GFI Slow water Fast water Fast water Fast water Dawn Dawn Noon Dusk 69 ± * 60 ± ± ± ± ± Table 2 Values of constants a and b used to describe the regression equations for the relationship between relative abundance of the three most common insects orders in stomachs and the fork length of the juvenile rainbow trout from Judge's Pool, Tongariro River (all samples combined; n = 74). * P < Order a r 2 Ephemeroptera Trichoptera Diptera * 0.24*

6 Dedual & Collier Food of juvenile rainbow trout 385 Taxon Fig. 3 Percentage composition of invertebrates found in the stomachs of juvenile rainbow trout caught at dawn ( h) in "fast" and "slow" water at Judge's Pool, Tongariro River, on 7 December 992. * relative abundance significantly different at P < E, Ephemeroptera; P, Plecoptera; T, Trichoptera; C, Coleoptera; D, Diptera; O, Other; Tr, Terrestrial. Taxon Fig. 4 Percentage composition of invertebrates found in the stomachs of juvenile rainbow trout caught in "fast" water at dawn (0645 h), noon (200 h), and dusk (630 h) at Judge's Pool, Tongariro River, on 7 December 992. Taxa as in Fig. 3. correlation existed between drift and trout stomach composition than between benthos and trout stomach composition (Z = 2.87; Z crit =.960; 0.00 < P < ). The relationship between the fork length of juvenile trout and the percentage of the three major insects orders found in their stomachs was described by a regression equation of the form: % Insect order = a FL + b where a and b are constants. Significant r 2 values described the relationship between fork length of the fish and the proportion of Trichoptera (r 2 = 0.42; P = 0.000) or Diptera (r 2 = 0.24; P = 0.03) in their stomachs (Table 2). Prey selectivity Pearre's (982) prey selection indices are presented in Table 3. In the drift, total Trichoptera had the highest positive prey selection index followed by adult caddisfly, A. neozealandica, total Ephemeroptera (notably Deleatidium spp.), and Maoridiamesa sp. fv > 0.5). Highest selectivity based on the benthic data was for unidentified Chironomidae, followed by Ephemeroptera (notably Deleatidium) and total Trichoptera (V > 0.0). In general, prey selection indices for the same taxa were higher for drift than for benthic samples. Because of the high abundance of Oligocheata in both the benthos and drift and their absence in the juvenile trout stomach samples, the V value for this taxon was highly negative (V = to-0.) (Table 3). DISCUSSION Juvenile trout diet The diet of juvenile rainbow trout collected from Judge's Pool during the day in December 992

7 386 New Zealand Journal of Marine and Freshwater Research, 995, Vol. 29 was dominated numerically by Diptera, followed by Trichoptera and Ephemeroptera. The importance of Diptera in the diet of the juvenile trout during summer has also been reported by Stephens (989) at another site on the lower Tongariro River. Ephemeroptera have been reported to be a major food source for juvenile trout elsewhere within New Zealand (Fechney 988; Glova& Sagar 99; Kusabs & Swales 99 ) and elsewhere (Angradi & Griffith 989), but in our study they represented only about 5% of the total prey items consumed. The benthic fauna of Tongariro River is considered typical of that found in other New Zealand rivers with gravel beds and moderate levels of nutrients (Quinn & Vickers 992). The similar pattern of invertebrate relative abundance in the benthos and drift suggests that most taxa were drifting in proportion to their abundance. The stronger correlation detected beween relative abundances of invertebrates in trout stomachs and in the drift compared with that in the benthos, provides further evidence that juvenile rainbow trout mainly feed on drifting organisms, as demonstrated by Elliott (973). Terrestrial organisms did not constitute a major proportion of invertebrates consumed by juvenile rainbow trout in our study, but in other New Zealand Table 3 Prey-selection indices as calculated from the number of prey in the stomachs of juvenile rainbow trout in relation to numbers of invertebrates found in the benthos and drift of Judge's Pool, Tongariro River on 7 December 992. ***, P < 0.00; *, P < 0.05 for X 2 -test. NA, not applicable. Taxon Ephemeroptera Deleatidium spp. Coloburiscus humeralis Plecoptera Zelandobius furcillatus Trichoptera Aotepsyche spp. Oxyethira albiceps Hydrobiosidae Adult caddisfly Diptera Cricotopus spp. Maoridiamesa sp. Chironomidae indet. Aphrophila neozelandica Adult Tipulidae Other adult Diptera Oligochaeta Terrestrial organisms Benthos 0. *** 0.2*** *** 0.08 *** 0.0*** *** 0.07 *** NA 0.0 * 0.02 *** *** 0.07 *** NA NA -0. *** NA Drift 0.7 *** 0.6*** *** NA 0.23 *** 0.08 *** 0.05 *** 0.0*** 0.9 *** *** 0.04* 0.8*** *** *** 0.00 streams they have been shown to be selected by juvenile trout (e.g., Glova & Sagar 99). Ellis & Gowing (957) and Peddley & Jones (978) suggested that high abundance of the exogenous fauna in the guts of some salmonids could reflect the scarcity of other prey in the river; this did not appear to be the situation on our December sampling date in the Tongariro River. Microhabitat effects Gut fullness indices were not significantly different between "fast" and "slow" water microhabitats at dawn, suggesting that fish were either feeding to similar extents in each habitat type or were returning to slow-water areas to rest after feeding. The average size of the juvenile trout at dawn in "slow" water was greater than in "fast" water. This observation contrasts with several other studies which found that salmonids move into swifter and deeper water as they grow (e.g., Everest & Chapman 972; Sheppard & Johnson 985). However, this may reflect movement of juvenile rainbow trout from feeding to resting stations, as has been observed in the Hinemaiaia River nearby (Dedual unpubl. data). The higher proportions of Ephemeroptera and Trichoptera and the lower proportions of Diptera found in the stomachs of juvenile trout from "slow" compared with "fast" water at dawn could be explained by the fact that the average size offish in the former sample was larger. We detected positive and negative shifts in the proportions of Trichoptera and Diptera, respectively, in trout stomachs with increasing fish length. The same shift has been described elsewhere in Tongariro River by Stephens (989) who found smaller proportions of Trichoptera and higher proportions of Diptera in fish smaller than 35 mm compared with fish longer than 50 mm. These observations support the hypothesis of Larkin et al. (957) that larger prey must be available for rainbow trout to increase their ration size and thus continue to grow, and are in accordance with the optimal foraging technique theory for predators as described by Krebs (979). Trichoptera and Oligochaeta showed significant variations in abundance in relation to flow microhabitat. Oligochaeta were significantly more abundant in "slow" water, as observed by Jowett et al. ( 99 ) in other New Zealand rivers. Trichoptera were more abundant in the benthos in "fast" water, and this is believed to largely reflect increased abundances of filter-feeding Aoteapsyche spp. at higher water velocities (see also Jowett et al. 99 ;

8 Dedual & Collier Food of juvenile rainbow trout Collier 993a). Although the Diptera as a group showed no pattern in abundance in relation to flow microhabitat, some chironomid taxa are known to have distinct flow preferences (Collier 993b). In particular, larvae of Maoridiamesa sp. intongariro River are more common in faster water velocities (see also Appendix ). Thus the abundance of some preferred prey items in the benthos and probably their delivery rates in the drift are likely to be higher in faster water. Diel variations In general there appeared to be little difference in the relative abundances of the different prey groups in the stomachs of juvenile trout between the diel periods examined. This contrasts with the results of Angradi & Griffith (989), who found highest proportions of Trichoptera larvae in the stomachs of 0- and l+-year-old wild rainbow trout at dawn. North American studies suggest that most Trichoptera are taken at night from the substrate (Bisson 978), although work by Angradi & Griffith (989) suggests that wild rainbow trout generally do not feed at night in summer. The absence of any significant variation in mean gut fullness indices between the dawn, noon, and dusk periods indicates that none of these daylight periods was more favoured for feeding. The large standard errors associated with the mean gut fullness indices suggest that there was considerable heterogeneity in feeding behaviour among the fish. Individual feeding differences have been reported in several other studies (Egglishaw 967; Elliott 967; Bryan & Larkin 972), suggesting that juvenile trout do not all feed at the same time. It is, however, unknown whether this is because the fish can adjust to their individual food requirements or because competition in a dominance hierarchy (Fausch 984) favours the dominant fish. This heterogeneity of feeding behaviour could be amplified by the number of meals taken during the diel sampling period. In experiments with brown trout (Salmo trutta), Elliott (973) showed that water temperature can limit the number of meals per day. In his experiments, trout fed twice a day at water temperatures ranging from 9 to 3 C (cf. 2 to 3 C in Tongariro River in our study), and it is therefore likely that the juvenile trout had at least two meals over our sampling period. Prey selectivity Although Diptera were numerically the most common prey type consumed by juvenile trout in 387 the size range mm during the day, they were not the most preferred food item. Indeed, only Maoridiamesa sp. and Aphrophila neozelandica were highly selected (P < 0.00 ) based on numbers in the drift. This may be related to prey size (these were the largest Diptera recorded in trout guts) since dietary selectivity in juvenile rainbow trout is known to reflect size differences of prey organisms (Bisson 978). Trichoptera, which represented 4% of the number of the prey consumed (cf. about 70% for the Diptera), was the most preferred taxon drifting at the Tongariro River site in December 992. Intensive exploitation of Trichoptera by salmonids has been reported previously (Egglishaw 967; Bisson 978; Peddley & Jones 978; Neveu 98), and prey size may again be a factor explaining this. Adult caddisflies and Hydrobiosidae were the most highly selected Trichoptera, and the comparatively low degree of sclerotised and inorganic material associated with them (compared with stony cased caddisfly larvae, for example) may account for their selection. Furthermore, adult caddisflies have a high fat content (Elliott 99) and for most fish this is the major energy source (Weatherley 976). Under the energy maximisation premise, predators choose their diet to maximise net energy intake per unit foraging time (Charnov 976). Ephemeroptera were the next most highly selected invertebrate group, and this seemed mainly to reflect a preference for nymphs of the leptophlebiid Deleatidium spp. The oligoneuriid Coloburiscus humeralis has been reported as the dominant prey consumed by adult rainbow trout in rivers elsewhere in New Zealand (e.g., McLennan & MacMillan 984). Although they were relatively common in the benthos of Tongariro River, few were apparently drifting during our sampling period. Furthermore, the high degree of sclerotisation and the large bifid gills of the nymphs may discourage juvenile trout from actively selecting C. humeralis. The apparent avoidance of Oligochaeta by juvenile trout could reflect the small size and possibly the opacity of these organisms, making visual detection difficult (e.g., Ware 972). Moreover, if they had been ingested, the absence of any sclerotised structures in their anatomy would make their detection in the stomach difficult. However, if this was so, we might have expected to have found some recently ingested Oligochaeta in the foreguts since oligochaetes have been reported in the stomachs of juvenile rainbow trout

9 388 New Zealand Journal of Marine and Freshwater Research, 995, Vol. 29 by Angradi & Griffith (989). It therefore seems probable that juvenile trout were avoiding Oligochaeta, either actively because of sufficient food supply provided by the other types of prey, or passively because of detection difficulties. CONCLUSIONS Our study has demonstrated considerable spatial variation in the abundance of potential food supplies for juvenile rainbow trout in the Tongariro River at the time of sampling. Abundances of benthic invertebrates commonly found in the diet of trout were generally highest in moderate to fast flowing water. Food delivery rates in the drift are also likely to be greater in this type of water, suggesting that maintenance of such microhabitats is important for food production. Individual trout appeared to feed at different rates independent of the variations observed in potential food abundance. Although the continuous food supply in the drift and the selective feeding of juvenile rainbow trout suggest that an adequate food supply was available in the lower Tongariro at this time, the wide variation in individual feeding rates could be interpreted as indicating a shortage of food. However, there are no accurate data available on the summer growth rates and densities of juvenile rainbow trout in the lower Tongariro River which could verify which of these scenarios is occurring. ACKNOWLEDGMENTS We thank the Fishery Team staff of the Department of Conservation for their assistance in the field, R. A. Faragher and an anonymous referee for perceptive comments on the first draft of this paper, and Dr Russel Cole from Waikato University for his statistical advice. REFERENCES Angradi, T. R.; Griffith, J. S. 989: Diel feeding chronology and diet selection of rainbow trout (Oncorhynchus mykiss) in the Henry's fork of the Snake River, Idaho. Canadian journal of fisheries and aquatic sciences 47: Bisson, P. A. 978: Diel food selection by two sizes of rainbow trout (Salmo gairdneri) in an experimental stream. Journal of the Fisheries Research Board of Canada 35: Bjornn, T. C. 97: Trout and salmon movements in two Idaho streams as related to temperature, food, stream flow, cover and population density. Transactions of the American Fisheries Society 00: Bryan, J. E.; Larkin, P. A. 972: Food specialization by individual trout. Journal of the Fisheries Research Board of Canada 29: Chapman, D. W. 966: Food and space as regulators of salmonid populations in streams. The American naturalist 00: Charnov, E. L. 976: Optimal foraging: attack strategy of a mantid. The American naturalist 0: 4-5. Collier, K. 993a: Flow preferences of aquatic invertebrates in the Tongariro River. New Zealand Department of Conservation, Science and Research Directorate, Wellington, New Zealand. Science and research series p. Collier, K. 993b: Flow preferences of larval Chironomidae (Diptera) in Tongariro River, New Zealand. New Zealand journal of marine and freshwater research 27: Cryer, M. 99: Lake Taupo trout production: A four year study of the rainbow trout fishery of Lake Taupo, New Zealand. New Zealand Department of Conservation, Science and Research Directorate, Wellington, New Zealand. Science and research series p. Egglishaw, H. J. 967: The food, growth and population structure of salmon and trout in two streams of the Scottish Highlands. Freshwater salmon fisheries research 38: -32. Elliott, J. M. 967: The food of trout (Salmo trutta) in a Dartmoor stream. Journal of applied ecology 4: Elliott, J. M. 973: The food of brown and rainbow trout (Salmo trutta and 5. gairdneri) in relation to the abundance of drifting invertebrates in a mountain stream. Oecologia 2: Elliott, J. M. 99: Rates of gastric evacuation in piscivorous brown trout, Salmo trutta. Freshwater biology 25: Ellis, R. J.; Gowing, H. 957: Relationship between food supply and condition of wild brown trout, Salmo trutta L., in a Michigan stream. Limnology and oceanography 2: Everest, F. H.; Chapman, D. W Habitat selection and spatial interaction by juvenile chinook salmon and steelhead trout in two Idaho streams. Journal of the Fisheries Research Board of Canada 29: Fausch, K. D. 984: Profitable stream positions for salmonids: relating specific growth rate to net energy gain. Canadian journal of zoology 62: Fechney, L. R. 988: The summer diet of brook trout (Salvelinus fontinalis) in a South Island highcountry stream. New Zealand journal of marine and freshwater research 22:

10 Dedual & Collier Food of juvenile rainbow trout 389 Glova, G. P.; Sagar, P. M. 99: Dietary and spatial overlap between stream populations of a native and two introduced fish species in New Zealand. Australian journal of marine and freshwater research 42: Hartman, G. F. 963: Observations on behavior of juvenile brown trout in a stream aquarium during winter and spring. Journal of Fisheries Research Board of Canada 20: Jowett, I. G.; Richardson, J.; Biggs, B. J. F.; Hickey, C. W.;Quinn, J.M. 99: Microhabitat preferences of benthic invertebrates and the development of generalised Deleatidium spp. habitat suitability curves, applied to four New Zealand rivers. New Zealand journal of marine and freshwater research 25: Krebs, J. R. 979: Optimal foraging: decision rules for predators In: Krebs, J. R.; Davies, N. B. ed. Behavioural ecology an evolutionary approach. Oxford, Blackwell Scientific Publications, pp Kusabs, I. A.; Swales, S. 99: Diet and food resource partitioning in koaro, Galaxias brevipinnis (Gunther), and juvenile rainbow trout, Oncorhynchus mykiss (Richardson), in two Taupo streams, New Zealand. New Zealand journal of marine and freshwater research 25: Larkin, P. A.; Terpenning, J. G.; Parker, R. R. 957: Size as a determinant of growth in rainbow trout, Salmo gairdneri. Transactions of the American Fisheries Society 86: McLennan, J. A.; MacMillan, B. W. H. 984: The food of rainbow and brown trout in the Mohaka and other rivers of Hawke's Bay, New Zealand. New Zealand journal of marine and freshwater research 8: Neveu, N. 98: Les rythmes alimentaires en milieu naturel. In: M. Fontaine ed. Nutrition des poissons. Coll. CNERMA 979, Paris. Pearre, S. Jr. 982: Estimating prey preferences by predators: uses of various indices, and a proposal of another index based on χ 2. Canadian journal of fisheries and aquatic sciences 39: Peddley, R. B.; Jones, J. W. 978: The comparative feeding behaviour of brown trout, Salmo trutta L., and the Atlantic salmon, Salmo salar L., in Llyn Dwythwch, Wales. Journal offish biology 2: Pitkethley, R. J. 990: Population dynamics of juvenile trout in two tributary streams of the Tongariro River. Unpublished MSc thesis. University of Waikato, Hamilton, New Zealand. Quinn, J. M.; Vickers, M. L. 992: Benthic invertebrates and related habitat factors in the Tongariro River. Consultancy report 6025/2, Water Quality Centre, Hamilton. Stephens, R. T. T. 989: Flow management in the Tongariro River. New Zealand Department of Conservation, Science and Research Directorate, Wellington. Science and research series 6. 5 p. Sheppard, J. D.; Johnson J. H. 985: Probability-of-use for depth, velocity, and substrate by subyearling coho salmon and steelhead in Lake Ontario tributary streams. North American journal of fisheries management 5: Ware, D. M. 972: Predation by rainbow trout (Salmo gairdneri): the influence of hunger, prey density, and prey size. Journal of Fisheries Research Board of Canada 29: Weatherley, A. H. 976: Factors affecting maximization of fish growth. Journal of Fisheries Research Board of Canada 33: Wurtsbaugh, W. A.; Li. H. 985: Diel migrations of a zooplanktivorous fish (Menidia beryllina) in relation to the distribution of its prey in a large eutrophic lake. Limnology and oceanography 30: Zar, J. H. 984: Biostatistical analysis (2nd edition). Prentice-Hall, Inc., Englewood Cliffs, New Jersey. 78 p. Appendix and 2 on following pages

11 390 New Zealand Journal of Marine and Freshwater Research, 995, Vol. 29 Appendix Numbers of invertebrates collected from the benthos in 0. m 2 Surber samples ( x ± SE; n = 5) in "slow" ( m s' ), "medium" ( m s" ) and "fast" ( m s" ) water habitats, and from the drift (no. in 2.5 h sample, n = ) at dawn ( h NZST), noon ( h) and dusk ( h) at Judge's Pool, Tongariro River on 7 December 992. For benthic data, standard errors < are not shown. Benthos Taxon Slow Medium Fast Ephemeroptera 8 ± Deleatidium spp Zephlebia sp. 2 - Coloburiscus humeralis 7 ± Plecoptera 2 2 Zelandobius furcillatus 2 Acroperla trivacuata - - Halticoperla viridans Spaniocerca sp. - - Trichoptera 55 ± Aoteapsyche spp. 8 ± 5 30 ± 4 Oxyethira albiceps ± Paroxyethira hendersoni - Psilochorema sp. 4 - Hydrobiosidae 8+ 9 ± 2 Helicopsyche sp. Pycnocentrodes spp ±9 Beraeoptera roria - Olinga feredayi - Pycnocentria funerea - P.evecta - Adult caddisfly - - Coleoptera Elmidae Hydrophilidae - Ptylodactylidae - Diptera ± 288 Cricotopus spp. 207 ± Eukiefferiella sp ± 7 Orthocladiinae sp. a 38 ± ± 49 Tanytarsus vespertinus Maoridiamesa sp. 464 ± ± 38 Chironomidae indet ±7 Adult Chironomidae - - Ceratopogonidae - Empididaesp. 0 ±4 3 + Muscidae sp. 2 ± 0 3 ± Aphrophila neozelandica 49 ± 76 ± 4 Eriopterini sp. - Austrosimulium sp. - Ephydrella sp. Adult Tipulidae Megaloptera 5 ± Archichauliodes diversus 5 ± 2 3 ± Oligochaeta 974 ± ± 23 Terrestrial - - Manuka beetle (adult) - - Other adult beetle - - Beetle larvae - - Adult Diptera - - Adult Hymenoptera - - Ant Unknown - - TOTALS 2760 ± ± ±9 9±7 0± ±56 3±2 6±3 3± ± ±50 98 ±44 44 ± ± ± 8 2± 2± ±475 Drift Dawn Midday Dusk

12 Dedual & Collier Food of juvenile rainbow trout 39 Appendix 2 Mean number (± SE) of organisms found in the stomachs of juvenile rainbow trout caught in different habitats and at different periods of the day from Judge's Pool, Tongariro River on 7 December 992. SWDA = "slow" water at dawn (0645 h); FWDA = "fast" water at dawn; FWNO = "fast" water at noon (200 h); FWDU = "fast" water at dusk (830 h). Taxon SWDA (n = 20) FWDA (n=7) FWNO(n= 7) FWDU (n = 20) Ephemeroptera Deleatidium spp. Zephlebia sp. Coloburiscus humeralis Plecoptera Zelandobius furcillatus Trichoptera Aoteapsyche spp. Oxyethira albiceps Hydrobiosidae Adult caddisfly Coleoptera Helodidae Hydraenidae Diptera Cricotopus spp. Eukiefferiella sp. Orthocladiinae sp. a Tanytarsus vespertinus Maoridiamesa sp. Chironomidae indet. Adult Chironomidae Empididae sp. Muscidae sp. Aphrophila neozelandica Adult Tipulidae Terrestrial Manuka beetle (adult) Other adult beetle Adult Diptera Unknown Fish TOTALS 2.50 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±.48