Phosphorus limited community dynamics of stream benthic algae andinsectsbyDarcie L. QuainmeB.Sc., The University of British Columbia, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of ZoologyWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAMarch 1994© Darcie L. Quamme, 1994In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of JO -1--L.The University of British ColumbiaVancouver, CanadaDate I&CC 3O/4DE.6 (2188)IIABSTRACTThe relationship between external soluble phosphorus (P) concentration and the abundanceand taxonomic composition of stream insects was determined in streamside artificial troughs.The response of peak algal biomass (PB) to target P concentrations of 0, 0.5, 2.5, 5, 10, 50j.g P 1 was also monitored. A log-linear function of P concentration was used toapproximate PB and insect abundance. PB measured as chlorophyll a increased with Pconcentration linearly to 7.4 mg m2 at 2.5 jg P 11 and reached an asymptote at9.2 mg m2 (2.7X the controls). Adult baetid mayflies showed a significant increase innumber after 23 days of P addition; this effect was maintained over 9 weeks of treatment.Numbers of benthic baetids, nemourid and perlodid stoneflies and hydroptilid tricopteranssampled at the end of the experiment significantly increased with P concentration. Adult andbenthic insects of these taxa exhibited similar rapid increases in abundance from 0 -2.5 g P. 1.1 and all showed signs of saturation at approximately 1.5 - 3 times the controls atconcentrations greater than 2.5 P l1. Increased abundances of insects resulted fromgreater food availability. There was no detectable difference in the numbers of large Baetidnymphs drifting from the troughs with increasing P concentration. Increased survival ofBaetidae nymphs with increasing P level was thought to account for higher numbers of adultand benthic baetids observed with increasing P concentration. Graphical comparisonsbetween the control and treated troughs showed that they were similar in taxonomiccomposition of insects. Insect taxonomic richness did not change with increasing Pconcentration. These findings are important to fisheries researchers who are assessing thepotential of stream fertilization as a technique to enhance salmonid populations in nutrientdeficient streams.H’TABLE OF CONTENTS1.0 . . . . . . . . . .... .. . .. . .. . . . . . . . . . . ...... ii2.0 LIST OF I’4B1FS .......... iv3.0 I_4ST OF IGiTRFS v4.0 ix5.0 16.0 OF SITE . 77.0 . 108.0 ......................................... 229.0 IISDLJSSION .. ... 5410. 7011. APPENDL’C 1 77lv2.0 LIST OF TABLESTable 1. Mean inorganic phosphorus concentrations (jtg P 1’) measured for eachtarget treatment ... 23Table 2. Mean inorganic nitrogen concentrations (g N 11) measured for eachtarget treatjnent..................... ....... . 24Table 3. Percent composition of the algal cell number by taxa of the artificialstream troughs for each target phosphorus concentration (ug P11).Percentages are based on algae collected from trough gravel atthe end of the experiment (July 8) .................... 31Table 4. Linear regression analyses relating total numbers (y) of variousinsect taxa collected as adults from the emergence traps totarget phosphorus concentration. Sample size (n) = 15 in all cases.Ordered with respect to decreasing p value.... . . . . . . . . . .34Table 5. Mean number of adult insects immigrating and emigrating pertrough day by taxon. Sample size (n) = 3. Standard error is given inparentheses . ..38Table 6. Linear regression analyses relating the numbers (y) of variousinsect taxa collected from the trough benthos at the end of theexperiment to target phosphorus concentration. Sample size (n) = 13in all cases. Ordered with respect to decreasing p value.49V3.0 LIST OF FIGURESFigure 1. Keogh River watershed with the location of the artificial streamtroughs 9Figure 2. Photo of mesocosm apparatus. The experimental treatments of thetroughs from left to right were 0.5, 0 + no insect immigration, 0,0, 2.5, 5, 10, 2.5, 5, 10, 0.5, 2.5, 5, 0, 0.5, 10, 50 jg• .. . . • •.. . • • . . •..• .. . . 13Figure 3. Mean algal biomass accumulation measured as chlorophyll afrom styrofoam plates. Sample size (n) = 3 except at 50 ugP11 where n = 1. Nutrient treatments started May 18.(3.1) First set of styrofoam plates, placed in troughs from April -May 31.(3.2) Second set of styrofoam plates, placed in troughs from May31 - July 8(3.3) Algal biomass accumulation (chlorophyll a) on the second setof styrofoam plates. The grazer excluded trough with nophosphorus addition (n = 1) compared to the means of thecontrol troughs (n = 3) 26Figure 4. Peak biomass on styrofoam plates measured as chlorophyll a(mgm2)against target phosphorus concentration. Linearregression of the data fitted by least squares. Regression lineabove (all data) is y = 4.05 + 419 log(P). r2 = 0.67. p < 0.001.Regression line (not shown) for treatments 0 - 10 ug Pt’ only isy = 5.09 * log(P) + 3.70. r2 = 0.69 p < 0.001 .... 27Figure 5. Density (cellsinm2)of algal cells collected from the trough gravel atthe end of the experiment (July 8) against target phosphorusconcentration. Linear regressions of the data were fitted by leastsquares.(5.1) Total cell numbers. Regression line above (for all data) islog(y) = 3.56 + -0.95 * log(P). r2 = 0.44. p = .003. Regression linefor treatments 0 - 10 ug Ph1 only is log(y) = -0.73 * log(P) + 3.48.= 0.21. p = 0.05.(5.2) Achnanthes mimutissima. Regression line above (for all data)is y = 316.41 + -200.19 * log(P). r2 = 0.48. p = .001. Regressionline for treatments 0 - 10 ug Pt’ only isy = -221.96 * log(P) + 324.83. r2 = 0.41. p = 0.005.(5.3) Oscilk#oria sp. Regression line above (for all data) islog(y) = 3.36 + -1.39 * log(P). r2 = 0.45. p = .002. Regression linefor treatments 0 - 10 ug P1’ only is log(y) = -1.29 * log(P) + 3.32.= 0.29. p = 0.02 ..... 30viFigure 6. (6.1) Daily minimum and maximum stream temperatures measured atthe mesocosm head tank as a function of time.(6.2) Numbers of adult insects caught in emergence traps between May19 and July 4 by target phosphorus concentration (ugPt’). In eachcase, the mean number of adult insects is plotted per date. Sample size(n) = 3 except at 50 ug P11 where n = 1....... ... 33Figure 7. Percent composition of the predominant adult insect taxa collectedfrom emergence traps of the artificial stream troughs at twophosphorus concentrations. Percentages are based on total insectscollected from May 19 to July 4 and result from the mean of threereplicate troughsFigure 8. Number of adult Baetidae caught in emergence traps between May19 and July 4 by target phosphorus concentrations (ug Pt’). Ineach case, the mean number of adult insects is plotted per date.Sample size (n) = 3 except at 50 ug P1’ where n = 1..... 37Figure 9. Number of adult Baetidae captured at the end of the experiment asa function of target phosphorus concentration (ug P1’). Linearregression lines of the data were fitted by least squares.(9.1) Numbers caught in emergence traps on July 4. Regression lineabove (all data) is y = 29.19 + 33.46 * log(P). r2 = 0.34.p = 0.011. Regression line (not shown) for treatments 0 - 10 ug P1’only is y = 23.16 * log(P) + 49.03. r2 = 0.51 p = 0.002(9.2) Numbers caught in drift nets on July 1.Regression line above (for all data) is y = 11.32 + 21.43 * log(P).= 0.69. p < .001. Regression line for treatments 0 - 10 ug P1’only is y = 20.47 * log(P) + 10.27. r2 = 0.51 p = 0.001..................39Figure 10. Contribution of Ephemeroptera to total biomass of adults collectedin emergence traps on June 27 by target phosphorus concentration(ug P1—’) . ‘14)Figure 11. Adult Ephemeroptera biomass (total dry weight of all individuals)collected on June 27 as a function of target phosphorusconcentration (ug P1-’). Linear regression line of the data fitted byleast squares. Variances were not stabffized by transformation.Regression line above (all data) is y = 8.193 + 14.241 * log(P).= 0.48. p = 0.002.Regression line (not shown) for treatments 0 - 10 ug Pt’ only isy = 11.76 * log(P) + 9.15. r2 = 0.25 p = 0.03 41Figure 12. Percent composition of Ephemeroptera collected as adults at the endof the experiment on June 27 by target phosphorus concentration(ug Pt’) 42vuFigure 13. Numbers of adult insects by target phosphorus concentrationcollected in emergence traps. All insects were captured betweenMay 19 and July 4. In each case, the mean number of adult insectsnumbers is plotted per date. Sample size (n) = 3 except at50 ug Pt’ where n = 1.(13.1) Chironomidae. (13.2) Trichoptera. (13.3) Plecoptera... . . . .. . . . . . . . . . . . . . . 45Figure 14. Number of insects collected from the benthos at the end of theexperiment on July 8 as a function of target phosphorusconcentration (ug P1’). Linear regression lines of the data werefitted by least squares. See Table 7 for the results of regressionanalyses carried out on 0 - 10 ug Pt’ treatments for the same data.(14.1) Baetidae. y = 152.06 + 89.96 * log(P). r2 = 0.443. p = .006.(14.2) Nemouridae. y = 117.92 + 93.02 * log(P). r2 = 0.421. p =0.008(14.3) Perloclidae. y = 9.41 + 13.31 * log(P). r2 = 0.402. p = 0.01(14.4) Hydroptilidae. y = 13.95 + 10.26 * log(P). r’ = 0.16. p = 0.08Figure 15. Dominant insect families collected from the benthos of the artificialstream troughs on July 8 as a function of phosphorus concentration.Percentages are based on the mean of three replicate troughs.50Figure 16. (16.1) Daily immigration rates of large Baetidae nymphs to theartificial stream troughs over time. Immigration rates are based on24 hour drift samples taken at the trough inflow.(16.2) Daily emigration rates of mature Baetidae nymphs from theartificial stream troughs over time by target phosphorusconcentration (ug P1-’). Emigration rates are based on 3 day driftsamples collected from the trough outflow. In each case, the meannumber of Baetidae nymphs is plotted per date with the associatederror bars. Sample size (n) = 3, for all points. 53Figure 17. Relative peak algal biomass (PB: PB(max)), chlorophyll a, as afunction of phosphorus treatment (ug P1’). PB(max) is PB at50 ug Pt’. A. modified from Bothwell (1989). B. Present study.Sample size (n) = 3 at each treatment level except at 50 ug Pt’where ii = 1 57VillFigure 18. Number of adult Baetidae captured in emergence traps per weekfrom July 8 - August 22 as a function of target phosphorusconcentration (ug P1-’). Linear regression lines of the data werefitted by least squares. Regression line above (all data) islog4y) = 1.41 + 0.49 * log(P). r2 = 0.74. p < 0.001.Regression line (not shown) for treatments 0 - 10 ug Pt’only is log(y) = 1.41+ 0.48 * log(P). r2 = 0.62. p = .002. . . . . . . ... . 1 • • • . . . . . . . .. . . .. . . . .76Figure 19. Number of insects collected from the benthos at the end of thesecond experiment on August 22 as a function of target phosphorusconcentration (ug Pt’). Linear regression lines of the data werefitted by least squares.(19.1) Baetidae. y = 2088.56 + 1701.56 * log(P). r2 = 0.45. p .003.(19.2) Nemouridae. y = 339.45 + 252.29 * log(P). r2 = 0.37. p = .008(19.3) Perlodidae. log(y) = 1.07 + 0.52 * log(P). r2 = 0.29. p = .02• • • 11••• • • I.e...... •••• • • • • •• I.e...... .77lx4.0 ACKNOWLEDGEMENTSI would like to thank my advisory committee which included Dr. M.L. Bothwell, Dr.T.G. Northcote, Dr. J.D. MePhail, Dr. G.G.E. Scudder, Mr. P.A. Slaney and Dr. C.J.Walters for advice and constructive comments on my thesis. I, also, appreciate the helpfulreviews provided by Dr. N.T. Johnston and Dr. H.A. Quamme. The following people wereimportant in the support of the project; Mr. K.I. Ashley, Mr. A.D. Martin, Mr. C.J.Perrin, Mr. B.O. Rublee, Mr. B.R. Ward and Mr. K.W. Wilcox. Also, Dr. J.H. Mundieand Dr. J.R. Post provided valuable suggestions on experimental design and techniques.This work was supported by an NSERC Operating Grant to Dr. T.G. Northcote and fundingfrom the British Columbia Ministry of Environment, Fisheries Branch and the NechakoFisheries Conservation Program. Valuable technical assistance was provided by Mark Beere,Jeff Burrows, Ray Carriere, Vera Bredow, Chris Giroux, Franzika Gross, Lindy Looy,Heather Quamme, Stephan Schug, Ryan Slaney, Katharine Staiger-Williams, Lynnel Steinke,Michelle Venne and the staff of the North Vancouver Island Salmonid Enhancement Society.Thanks to T.W. Chamberlin and the staff of the B.C. Ministry of Environment, Fish andWildlife Branch, Skeena Region Headquarters for work space and support provided inSmithers, British Columbia.1Phosphorus limited community dynamics of stream benthic algae andinsects5.0 INTRODUCTIONThe relative importance of predators and resources in controlling the abundance or biomassof aquatic organisms has been long studied in lakes and ponds (e.g. Hrbacek et al. 1961,Arruda 1979, Shapiro 1979, Shapiro and Wright 1984, Carpenter et al. 1985, 1987, Ranta etal. 1987, Perrson et al. 1988). About half of the explained variation in productivity amonglakes is thought to result from nutrient supply, turnover time of the water and vertical mixing(Schindler 1978, Schindler et al. 1978, Carpenter and Kitchell 1987, 1988). The remainderresults from the effects of predation “cascading” through to lower trophic levels (Carpenter etal. 1985, Carpenter and Kitchell 1988).In contrast, experimental assessments of bottom-up (resource limitation) and top-down(predation) trophic interactions in riverine and stream ecosystems are relatively few (Deeganand Peterson 1992, Power 1984b, Peterson et al. 1993a, Peterson er al. 1985, Feminella etal. 1989, Johnston et al. 1990). The extent to which bottom-up nutrient control in streamspropagates throughout the food web has been poorly studied.Recently, studies have shown that the biomass of insects (Hart and Robinson 1990,Mundie et al. 1991) and insectivorous fish (Slaney and Ward 1993, Peterson et al. 1990)2may be enhanced by addition of inorganic phosphorus and nitrogen through increases inautotrophic production.Some periphytic algae are thought to be high quality food for many aquatic insect grazers(Lamberti and Moore 1984). Nitrogen and phosphorus augmentation can enhance arealbiomass of benthic algae (Stockner and Shortreed 1978, Elwood er a!. 1981, Grimm andFisher 1986, Perrin et al. 1987, Bothwell 1989, Lohman et a!. 1991) and microbial growthin streams (Peterson et a!. 1985, Hullar and Vestal 1989). The effect of external phosphorusconcentration on algal biomass (periphytic diatoms) has been well described by Bothwell(1989).The abundance or biomass of insect algal grazers in streams may be set by the availabilityof periphytic algae. It is though that the stream invertebrates are food limited (Richardson1993). Evidence to support this hypothesis comes from correlative studies betweenmacroinvertebrate densities and their food supply (Egglishaw 1964, McKay and KaIf 1969,Hawkins and Sedell 1981, Drake 1984, Flecker 1984, Murphy et al. 1986, Richardson andMcKay 1991), invertebrate density manipulations (Lamberti et al. 1987, McAuliffe 1984a, b,Hart 1985, 87, Richardson 1993) and food supply manipulations.The extent to which nutrient levels affect different elements of the periphytic communityand translate into population effects for aquatic insects has rarely been determined.Information concerning insect response to nutrient enhancement in terms of abundance,composition, and timing of response is limited (Mundie et al. 1991). The relationship3between ambient nuthent concentration and the insect abundance, biomass and compositionhas never been quantified experimentally in streams over a broad range of nutrient levels.Previous fertilization experiments have usually involved a single treatment at a high levelof nutrients in either artificial stream troughs or whole rivers. Mundie et a!. (1991)assessed the insect community response to nutrient enrichment for one level of phosphorus(10 ug P11) in fish-excluded flow-through troughs and found increases in total insectnumbers mainly due to higher numbers of Chironomidae. Hart and Robinson (1990)followed the effect of phosphorus addition at one high level (60 ug Pl1) on two species ofgrazing caddisflies, Leucotrichia pictipes and Psychomyia flavida, in wooden flumes wherefish, crayfish and predatory insects were present. They found that the larvae of both insectspecies had higher individual mass, developmental rates, and population densities inphosphorus treated flumes.Three whole river fertilization studies (Elwood et al. 1981, Johnston et a!. 1990, Petersonet a!. 1993a) have resulted in greater abundance or biomass at higher trophic levels includinginvertebrates and/or fish. Increased densities of the snail, Goniobasis clavaeformis, werefound in an experimentally enriched reach of Walker Branch, a Tennessee woodland stream(Elwood et a!. 1981). Johnston et al. (1990) and Deegan and Peterson (1992) found thatfertilization resulted in greater fish growth. Increases in sizes of fish have been attributed toincreased algal productivity and higher insect abundances, a major component of fish diet.On the Keogh River, in British Columbia, mean weights of juvenile steelhead trout and4coho salmon increased with fertilization (Johnston et al. 1990). A preliminary study, by thesame researchers, suggested that the upper Keogh River produced 3-5 times the total insectbiomass in fertilized sections than untreated sections; however, the effects of nutrientadditions could not be separated from location effects. The authors suggested that increasesin fish biomass may have resulted from greater insect abundance with nutrient treatment.Their proposition was supported by previous studies in which lotic salmonid abundance(Slaney and Northcote 1974, Mason 1976, Wilzbach et al. 1986) and size (Mason 1976,Wilzbach et a!. 1986) had been shown to increase with food availability. Salmonid standingstocks (Murphy et al. 1981, Bowiby and Roff 1986) and growth rates (Warren et al. 1964,Wilzbach et al. 1986) are often correlated with biomass of benthic prey. Researchers havealso found that fish growth (Gibson and Haedrich 1988) or standing crop (Hoyer andCanfield 1991) is correlated with nutrient levels measured in stream surveys.Whole river fertilization increased the size of both young of the year and adults of Arcticgrayling on the Kupuruk River in Alaska (Peterson et al. 1993a). Neutral lipid storage inadult grayling was increased in treated areas compared to controls (Deegan and Peterson1992). The abundance of the total number of insects was not strongly affected byphosphorus additions but the relative importance of different species shifted over time.There were increases in numbers of Baetis lapponicus and Brachycentrus americanus whileProsimulium martini declined in abundance (Peterson et al. 1993a). The growth rates of thefour dominant large insect species increased in response to treatment. By examining thenitrogen and carbon stable isotope ratios of Kuparuk River food web components, Peterson etal. (1993a) traced algal production stimulated by fertilization through to tissues of insect5ivorus grayling.It has been suggested that fertilization may be used as a technique to enhance streamrearing salmonid populations that are food limited (Johnston et al. 1990). Whole riverfertilization has been shown to increase smolt size and may improve marine survival ofstream reared coho and steelhead (Johnston et a!. 1990 and Slaney and Ward 1993).Increased adult returns have been reported for coho salmon (Hager and Nobel 1976) andsteethead trout (Ward and Slaney 1988) with greater smolt size.Despite the body of knowledge described above, few studies of streams have examined theeffect of nutrient concentration on the abundance or biomass of higher trophic levels under auniform set of environmental conditions. The objectives of this study were to quantativelyestablish the relationships between: (1) phosphorus concentration and the areal biomass ofperiphytic algae, and (2) phosphorus concentration and the abundance and composition ofstream insects. Uniform physical conditions were established by carrying out experiments ina stream mesocosm, 17 artificial stream troughs located on the Keogh River. This river nearPort Hardy, British Columbia, was selected as a study site because it is a nutrient limitedcoastal stream. Previous research involving a whole river fertilization experiment conductedon the Keogh River between 1983 and 1986 (Johnston et al. 1990), provided backgroundexperimental data. Experiments were simplified by excluding insectivorous fish from themesocosm.Previous work has shown that phosphorus is very low (< 1 g P. l’) in the Keogh Riverand suggested that phosphorus may be the primary limiting nutrient to productivity (Perrin et6al. 1987). Target nutrient concentrations for this study (0.5 - 50 g P 1.1) were selected toenhance growth rates of cells within the algal mat that might be phosphorus limited due toslow diffusion of phosphorus to the underlying cells (Bothwell 1989). Nitrogen was added ata constant ratio to ensure that it did not become limiting.The establishment of the above relationships is important to fisheries researchers who areassessing the potential for fertilization of nutrient deficient coastal streams as a technique toenhance salmonid populations in British Columbia (Johnston et al. 1990, Slaney and Ward1993). This study will also provide information for aquatic resource managers concernedwith riverine eutrophication.76.0 DESCRIPTION OF STUDY SITEThe experiment was carried out on the Keogh River (127’25’ W by 50°35’N), a thirdorder coastal stream on northeastern Vancouver Island, near the town of Port Hardy insouthwestern British Columbia. The experimental mesocosm was located 31 km upstream ofthe river mouth, approximately 1 km downstream from Keogh Lake outlet (see Figure 1).The river is 32 km long with a 130 km2 watershed in the Coastal Western Hemlock forest.Western hemlock (Tsuga heterophylla), red cedar (Thuja plicata), and spruce (Piceasitchensis) dominate mature forests. Approximately 45% of the watershed has been loggedduring the last 30 years. The headwaters of the Keogh River were logged to thestreambank about 20-25 years ago including the area where the experiment was located. Inlogged areas the major riparian species are red alder (Alnus rubra), salal (Gaultheriashallon), willow (Salix spp.), and sedges and grasses (Johnson et al. 1990) and to a lesserextent coniferous species. Detailed descriptions of the Keogh River are given in Ward andSlaney (1979), Slaney et al. (1986), Perrin et a!. (1987), and Johnston et a!. (1990).The wetted mean width of the river at mean summer flow is 8.1 m (range 5-15 m). Meanannual discharge is 5.3 m3s1 with a minimum in mid-summer of 0.1 m3s1 and an estimatedwinter maximum flow of 254m3s’. Mean summer flows in the upper-river (28 km from themouth) and near the mouth (3 km) are about 0.5 m3s’ and 1.6m3s’, respectively (valuesfrom Johnston et a!. 1990). In 1991, at this experimental site, the wetted width ranged from6.0 - 7.0 m and the flows ranged from 0.14 - 0.39 m3s’ between May 10 and8July 8.Ambient nutrient concentrations of the Keogh River in spring to summer are very low.Johnston et a!. (1990) determined soluble reactive phosphorus (SRP) levels to be < 1 gtotal dissolved phosphorus, 5 g l’; nitrate nitrogen, < 15 g 11; and total ammonia,< 5 g 1 . Nutrient concentrations have been found to elevate with declining flowsthrough the summer in the upper river (Perrin et al. 1987); inorganic dissolved phosphorusincreased slightly while nitrate nitrogen increased up to 4-fold. Mean pH is 6.9, totalalkalinity is 7.0 mg l’ and total dissolved solids are 30 mg i (Johnston et a!. 1990).Mean pH at the study site was 6.9; mean true colour, 19; mean turbidity, 1.3; and meantotal alkalinity, 9.0 mg l’ at the mesocosm head tank, based on five collection dates (1991;May 29, June 4, 10, 18, 25) at a mean flow of 0.24m3s1. The extremes of watertemperatures at the study site were 12.5 - 21 °C over the course of the experiment.9JJ- CI—F.— —--, —.‘ 4— “)O‘ MIsy Lake1’o 123 45ksi \ e—Scai.0 Connor• Lak.C MuirL_t27/Varcve( oQ. -%Iad Pott )EXPERIMENTAL )SITEIiLake /0 3Ok OG- L7 IRI’1E.Figure 1. Keogh River watershed with the location of the artificial streamtroughs.107.0 METhODS7.1 Description of mesocosm apparatusThe mesocosm consisted of seventeen flow-through troughs (each 1.52 m x 0.20 m x 0.20m) assembled at the stream side. The troughs were fabricated from clear plexiglass. Waterand biota from Keogh River were delivered to the mesocosm by gravity through a 300 mlong (15 cm diameter) plastic pipeline fitted to a head tank. Water flow from the head tankto each of the troughs was adjusted to approximately 1 l.51 by the use of rotating polyvinylchloride (PVC) standpipes and flexible nylon tubing similar to Mundie et al. (1991). Highrates of hydraulic flushing within the troughs (approximately 9 seconds residence time)ensured that trough water was similar in quality to river water and that added nutrientsremained constant over the trough length. Water flows in the troughs were checked andadjusted at least 3 times per week. Excess water, maintained in order to adjust water flow tothe troughs, overflowed from slots at the back of the head tank and a valve off the mainpipeline.Each trough was filled with gravel to a depth of 7 cm over an area of 0.30 m2. Numbersof lotic insect larvae have been found to be highest in the upper 10 cm of stream substrates(William and Hynes 1979). The design of the artificial streams in this experiment did notallow for vertical and horizontal migration of crawling insects to and from the troughs.Seventeen percent of trough sand and gravel passed through a seive size of size 7.9 mm;forty-three percent was 7.9 - 19.1 mm, and forty percent was 19.1 - 50.8 mm (by mass).Four baskets (100 cm2 each) filled with gravel were placed at the downstream end of the11gravel for purposes of subsampling benthic insects. Water in the troughs covered the gravelto a depth of 3.0 cm (standard error (SE) = 0.3, n = 17). Surface velocities of the troughsaveraged 21.3 cms’ (SE = 0.5, n = 17). Results of simple linear regression analysesdemonstrated that there was no systematic variation in surface velocity across the mesocosm(p = 0.34, r2 = 0.00, t6 = -0.98). The trough environment was designed to simulatethe characteristics of a stream riffle including a fast current, coarse substrates, some sandcontent and a low accumulation of organic matter. Water depths in the troughs were similarto only the shallowest areas of the natural stream riffles where water depths ranged 5 - 14cm and water velocities averaged 27.3 cms1 (for six sites located near the mesocosm).The bottom of the trough, downstream of the gravel, was fitted with a sheet (320 cm2) ofopen cell Styrofoam DB (Snowfoam Products Ltd., El Monte, California) that provided asubstrate for periphytic algae that could be easily monitored throughout the study.Artificial substrates such as styrofoam are commonly used to reduce periphyton variability.In a recent study by Morrin and Cattaneo (1992), chlorophyll estimates were found to besignificantly less variable on artificial than on natural substrates. However, other estimatesof biomass in this study (including ash-free dry mass, dry mass, density and biovolume) didnot confirm this finding. I used chlorophyll a per area as a means to estimate algal biomasson the styrofoam substrates. Styrofoam substrates selectively promote the colonization ofperiphytic diatom communities (for example Bothwell 1989). The mesocosm site waspartially shaded by a vascular plant canopy. All troughs received full shade, partial shadeand full sunlight during the course of a day as shadows moved across the mesocosm.12A funnel trap for emerging adult insects covered the length and completely sealed the topof each trough (Mundie et a!. 1991). Openings on the side walls of the traps were fittedwith 355 m mesh Nitex netting to allow air movement but, prevent the escape of capturedinsects. All adult insects were funneled into a drop trap containing 50% ethanol plus severaldrops of liquid soap to act as a surfactant to reduce surface tension in order to trap insectsmore easily.Nitex drift nets (100 m mesh) were used to monitor insect immigration to and emigrationand from the troughs by filtering the inflowing or outfiowing stream water. The inflow netsprevented insect immigration and colonization of the troughs one day a week during theexperimental period (described in Methods).13Figure 2. The mesocosm apparatus. The experimental treatments of thetroughs from left to right were 0.5, 0 + no insect immigration, 0,0, 2.5, 5, 10, 2.5, 5, 10, 0.5, 2.5, 5, 0, 0.5, 10, 50 g P11.147.2 Nutrient treatments and physical measurementsThis experiment involved six target levels of total phosphorus; 0, 0.5, 2.5, 5, 10,and 50 g P 11. Each treatment was replicated 3 times except for the 50 g 1 treatmentwhich was unreplicated. Treatments were assigned to the troughs at random so as not tocorrespond with possible effects resulting from horizontal gradients across the troughs.Nitrogen was added to troughs at a constant ratio of 2:1, N:P (atomic weight) or 0.9:1(wtlwt). Blooms of inedible blue-green algae can occur at low N:P ratios in lakes (Schindler1975, Tilman 1977). Previous whole river fertilization of the Keogh River at an N:P ratio of1:1 (wt./wt.) in 1984 did not result in a bloom of blue-green algae (monitored usingstyrofoam substrates). A similar N:P ratio was chosen for the present experiment.Background levels of dissolved inorganic nitrogen (< 20 g N• 1.1) on the Keogh Riverwere thought to be high enough to permit maximum algae cellular growth rates under naturalconditions. A bioassay conducted on the Nechako River (Perrin 1989) examined the biomasslevels of periphytic algae as a function of nitrogen addition (0 - 100 g N• l’) at surplusphosphorus concentrations. In Pen-ins’s experiment sixty to seventy percent of the maximumbiomass response measured was achieved at a N concentration of 10 g 1’.The source of nutrients (supplied by Coast - Agri, Abbotsford, British Columbia) used inthe present study was a liquid agricultural fertilizer blend of 32-0-0 (50% urea, 25%ammonium and 25% nitrate by mass) and 10-34-0 (10% ammonium and 34% totalphosphorus asP205). Total phosphorus was 100% water soluble and available as 25-35%ortho-phosphate and 65-75% 4-12 chain polyphosphates. The ammonium polyphosphate15fertilizer (10-34-0) was also comprised of micronutrients. These included Boron (0.01 % Bby mass), Calcium (0.07% CaO), Copper (0.0005% CuO), Iron (0.56% Fe203), Magnesium(0.5% MgO), Manganese (0.015 MgO), Potassium (0.12% K20), Sulfate (1.8% SO4) andZinc (0.10% ZnO). This fertilizer blend was used because it has been shown to be costefficient and easily applied to whole rivers in large scale studies (Slaney and Ward 1993).Beginning April 27, 1991 water from the stream was run through the troughs for 21 daysto allow colonization by stream biota. Nutrient additions began May 18 and ran until July 8,1991. A second experiment was carried out from July 8 - August 22. It was thought thatpolyphosphates in the agricultural fertilizer might not have been biologically available overthe length of the troughs. Nutrient levels in the second experiment were the same as the firstexperiment but fertilizer types were switched from the agricultural fertilizer to readilyavailable reagent grade nutrients in an attempt to achieve a more dramatic increase in algalbiomass with increasing phosphorus concentration. In the second experiment, Totaldissolved phosphorus was available as Soluble orthophosphate. The second experiment was,however, confounded by the first because gravel and insects from the first experiment werenot replaced. Styrofoam substrates for algal biomass accrual were replaced. Results of thesecond experiment are not reported in detail here.Fertilizer was dripped into the head of the artificial stream troughs via microbore tubingusing a Technicon auto-analyzer pump. Rate of nutrient delivery was adjusted by alteringmicrobore tubing size and stock concentration. Phosphate concentrations in troughs werecalculated from known dilutions of standard phosphorus and nitrogen solutions. A plexiglassbaffle at the head of each trough created water turbulence and ensured mixing of stream16water and nutrient additions.Samples for water chemistry were taken at weekly intervals from all troughs in order tocheck computed nutrient additions. Because of limited sensitivity of analytical methods forphosphorus, only higher levels of phosphorus additions (2 - 50 g P. 1’) could be accuratelyverified. Water samples were collected and analysed within 24 hours. Nitrogen andphosphorus analyses were carried out by Zenon laboratories, Bumaby, British Columbia on aTechnicon AA2 auto-analyzer. Analyses of NO3 + NO2 -N, N02-N, Total P, Soluble TotalP and Soluble ortho-P were performed according to modified procedures of Taras et al.(1971) described in McQuaker (1976). Analyses of NH4-Nwere performed according tomodified procedures of Greenberg (1980) described in McQuaker (1989). The limit ofdetection was 3 g P 1’ for Total P, Soluble Total P and Soluble ortho-P; the limit for lowlevel Soluble ortho-P was 1 g P 1’. The detection limit for NH4-N and N02-Nwas5 g N 1’; low level N02-Nwas 1 g N and NO3 + NO2 -N and N03-Nwas20 g Nl’.Water temperature was measured with a continuously recording Ryan model J-90submersible thermograph, placed in the mesocosm head tank. The instrument was calibratedweekly. No difference was found between stream temperature and water running through themesocosm.177.3 Algal communityAlgal biomass was sampled by removing two periphyton cores (6.2 cm2), weekly, fromstyrofoam sheets in each trough. The time-course of algal biomass accrual as measured bychlorophyll a was monitored on styrofoam plates from April 2 - May 31. These plates werereplaced when algae the sloughed from styrofoam, and new plates were monitored from May31 - July 8.Chlorophyll a analyses followed the procedure of Parsons et a!. (1984). The styrofoamcores were frozen at -20 °C until later (3-5 months) chlorophyll a flourometric determination.Algal cells were disrupted in 10-15 mL of 90% acetone (at 0°C) using a Potter-Elvehjemtissue grinder (3 minutes) and a sonication bath (5 minutes). The homogenate was incubatedin a dark refrigerator (3°C) for approximately 20 hours and then centrifuged for 5 minutes atsetting 5 in a Damon IEC clinical centrifuge to remove solids. Chlorophyll a from each corewas measured separately using a Turner Designs model 10 flourometer. A correction wasmade for phaeophytin (Parsons et al. 1984). Duplicate cores from each trough wereaveraged for each date. Bothwell (1983) found that chlorophyll a levels extracted fromindividual cores (5.0 cm2) of algae (n = 5) adhering to styrofoam had a coefficient ofvariation of 10.9%. In my study, duplicate cores from each trough had coefficients ofvariation that ranged from 2 - 55% (mean = 14%) across the seventeen troughs (determinedbefore treatment initiation, May 17).The peak algal biomass (PB) found on the styrofoam substrate over time was used to18describe the relationship between P concentration and areal algal biomass (Bothwell 1989).PB was estimated by averaging chlorophyll a values observed on the second set of styrofoamsheets (May 31 - July 8) for the final two collection dates (July 3 and 8). Peak chlorophyll alevels on the first set of styrofoam sheets occurred after only eight days of treatmentapplication and were thus not used to estimate PB.Extraction and quantification techniques of chlorophyll a used in this study were comparedto those used previously (Johnston et a!. 1990 and Perrin et a!. 1987). Four cores ofperiphyton were sampled per sheet of styrofoam with various levels of accrued periphytonbiomass from the Slocan River near Nelson, B.C. Eighteen styrofoam sheets were sampledin total. Chlorophyll a on two of the cores of each sheet was measured using techniques ofthis study described above and two were analyzed using previous techniques and were carriedout by Zenon laboratories.Periphyton adhering to gravel was sampled on July 8 and fixed in Lugol’s solution fortaxonomic identification and cell counts. Two stones per trough were dropped directly inLugol’s solution of a standard volume of 500 ml; later, algae was scraped from both stonesand pooled. Quantitative cell counts were made at 500X magnification in Utermohlchambers from subsample volumes of 25 ml. A minimum of 100 individuals of thepredominant species and at least 500 cells in total were counted. The areas of the rockswere calculated with a planimeter after obtaining an impression of the surface withaluminium foil.197.4 Insect communityAdult insects were collected weekly from emergence traps and preserved in 90% ethanol.Drift nets that collected emigrating insects from the outflow of the troughs were emptiedtwice a week (3 day collection periods). Once a week, drift nets were placed on the inflowof the troughs for 24 hours to assess the immigration of insects to each trough. One pipefrom the head tank was monitored for immigrating insects three times a week (two 3 day andone 24 hour collection period). As a result drift nets excluded insects from one troughfollowing the 21 day colonization period described above.Benthic insects from two of the baskets in each trough (200 cm2) were sampled at the endof the experiment, July 8 . Drift nets (100 jm mesh) were simultaneously placed on inflowand outflow of each trough while baskets full of gravel were removed from the trough.Insects were released from the gravel by gently brushing and rinsing the stones. Releasedinsects were captured in the outflow net. Drifting and benthic insects were preserved in 5%formalin.Benthic insects were also sampled directly from the stream using a Hess sampler of 0.05m2 with a 100 m mesh net. The substratum was sampled to a depth of approximately 10cm during base flow conditions, in water depths of 7 - 15 cm. Three samples were takenfrom three different riffles.Samples were hand-picked at lox magnification. Sorting efficiencies were >90% forbenthic and drifting insects determined by repeated picking. All adult insects were identified20and enumerated. Drifting and benthic insects were sieved through nested screens. Benthicinsects greater than 650 m were identified and enumerated. Generally, drifting insectsgreater than 1000 m were enumerated and identified. All sizes of drifting insects samplednear the beginning of the experiment (May 31 and June 3) were examined.Insect taxonomy follows Merritt and Cummins (1984). Invertebrates were identified atleast to the family level and occasionally to genus. Generally, insect taxa were not sub-sampled. However, a gridded petri dish was used to sub-sample drifting Baetidae. Varianceto mean ratio tests (x2 tests) for agreement with a Poisson series were performed accordingto Elliott (1977).Results (below) show that Ephemeroptera (mayflies) was the only order collected tocontain adult taxa that increased in number with increasing phosphorus treatment. Thus theeffect of treatment on Ephemeroptera adult body length and total biomass was assessed forinsects collected near the end of the experiment (June 27). Body length measurements weremade using a dissecting scope at lox magnification and a Summasketch Model MMIIIdigitizer calibrated with a stage micrometer. To make biomass measurements, insects wereremoved from ethanol, oven dried for approximately 20 hours at 60°C, and weighed on aCahn/Ventron 21 automatic electrobalance.217.5 StatisticsAll linear regression analyses were completed using Microsoft Excel version 4.0 by leastsquares fitting methods. In all tests, the critical level of significance was a = 0.05. Whenthe relationship between the variables was visibly non-linear, I did a log transformation ofthe x axis that linearized the data prior to the analyses (Zar 1984). A log transformation ofthe y axis was made occasionally in order to stabilize the variance. Assumptions of theregression were validated before and after transformation by graphical examination ofresiduals.All regressions were initially completed on replicated treatments only (0 - 10 jg P 1.1) totest for significance. These results are presented in tables and text. If the above regressionswere significant I carried out regression analyses of the same data, but also included the highunreplicated treatment 50 g P 1 and displayed the results graphically.228.0 RESULTS8.1 Nutrient treatments and Algal Community ResponseIncreases in phosphorus addition to troughs led to measured increases in Total P (Table 1).Measured Total phosphorus above background levels was approximately three times lowerthan target levels of phosphorus for treatments of 2.5-10 g P. 1*1. Measured nutrients wereelevated up to 8.3 and 2.6-3.75 times the controls for Total P and Dissolved InorganicNitrogen, respectively (Table 1 and 2). Measured N:Total P ratios ranged from 1.5 - 5 inexperimental and control troughs.There was no significant difference among troughs in chlorophyll a levels beforetreatments began (measured from cores collected on May 17, 45 days after initial placementof the first set of styrofoam plates in the troughs, p = 0.66, r2 = -0.06, t05)j4 = -0.44, fortreatment troughs 0-10 tg P. 1.1). There was no systematic effect of trough location onchlorophyll a levels across the mesocosm for the same date (p = 0.40, r2 = -0.02, t05 =-0.86). A significant increase in chlorophyll a with increasing phosphorus concentration wasobserved after only 6 days (May 24) of treatment (p = .001, r2 = 0.54, t05)14 = 4.05)(Figure 3.1). However, I was not able to follow the effects of phosphorus treatment on theaccrual of algal biomass on these styrofoam plates over time because algae began sloughingoff the plates.23Mean inorganic phosphorus concentrations (jig P 1’) measuredfor each target treatment.Target Measured Soluble SolubleTotal P Total P2 Total P2 ortho-P2Addition mean(SE) (range) (range)0 4.6(0.6) <3-3 -0C-)Figure 10.TA3ET PHOSPHORUSCONCENTRATION (ug P/I)414O. .Cl)cx<30DOz •.JLUDI-.uJujLL. Lii0ODO,LL 10Lo0- .,,,,, I0 .5. 2.5 5 10 50TARGET PHOSPHORUS CONCENTRATION (ug P/I)Figure 11. Adult Ephemeroptera biomass (total dry weight of all individuals)collected on June 27 as a function of target phosphorusconcentration (ug P1’). Linear regression line of the data fitted byleast squares. Variances were not stabilized by transformation.Regression line above (all data) is y = 8.193 ÷ 14.241 * log(P).r2 = 0.48. p = 0.002.Regression line (not shown) for treatments 0 - 10 ug P1’ only isy 11.76 * log(P) + 9.15. r2 = 0.25 p = 0.03.42LUL100ZWZOcLxI- LU0LU=I-Figure 12. Percent composition of Ephemeroptera collected as adults at theend of the experiment on June 27 by target phosphorusconcentration (ug PT’).100- IIHIIIIIIIIIIIIIIII—III HeptageneidaeParaleptophiebia sp.• Baetiso 10PHOSPHORUS CONCENTRATION(ug P/I)43t05(2)13 = -0.80, for treatments 0 - 10 pg P.1-i). Adult mayflies (collected near experimentend on July 4) exhibited no difference in mean body length per trough as a result oftreatment (p = 0.49, r2 = -0.04, t()13 = 0.71, for treatments 0 - 10 g Plj. Adult bodylengths ranged from 2.04 to 6.85 mm.In two taxa, Chironomidae and Tricoptera, adults captured in emergence traps increased innumber only at the highest nutrient concentration of 50 g P1’ (unreplicated trough).Increasing phosphorus concentration increased adult Chironomidae and Trichoptera numbersbut the effect was nonsignificant from 0-10 jg P-1’ (see Table 4 for regression statistics).However, at 50 g Pl the numbers of adult Chironomidae collected at the end of theexperiment (July 4) were three times the controls (Figure 13). The number of adultTrichoptera captured from the 50 g P1-1 trough was eight times that of the control troughsat peak emergence (June 27) (Figure 13). The numbers of adult Chironomidae collected indrift nets on trough outflows near the experiment end (June 24) showed no significantresponse to increasing phosphorus concentration (p = 0.92, r2 = -0.08, t05(2)13 = -0.10, fortreatments 0-10 g P14).The mean number of adult Plecoptera emerging per week declined over the course of theexperiment to less than one per week by July 4 for all treatments (Figure 13). No significanttreatment effects were observed on the total number of Plecoptera captured from May 19 toJuly 4 (see Table 4). Other major insect taxa collected in the emergence traps (Simuliidae,miscellaneous Diptera; Leptophlebiidae; Heptageniidae; and Coleoptera) showed nosignificant response to fertilization Table 4). The relationship of increasing phosphorusconcentration on the numbers of adults of taxa other than Baetidae and Chironomidae44collected in the drift near the end of the experiment (June 24) could not be assessed becauseof low numbers.A descriptive comparison was made between number of adults drifting into the troughs perday and those emigrating from the troughs per day for 0 and 10 P-i4 treatments. Thetotal number of adults immigrating per day to the troughs was based on a 24 hour driftsample. The total number of adults emigrating from the troughs per day was estimated byadding the daily catch of adults of the emergence trap (sampled over 10 days) to those in thedrift nets (sampled over 3 days). Greater than 91 % and 79% of the baetid and trichopteranadults trapped per day, respectively, originated from trough gravel for these treatments(Table 5). None of the simuliid nor chironomid adults trapped per day originated fromtrough gravel (Table 5). Comparisons for other taxa were not made because of lownumbers.4513.1 13.2250 80E -- .EgI200 ‘III.’I SI SaILdISO40 ‘‘U100ZO4Z 2O50 uacJI-010 17 23 30 40 *7 0 0 10 17 23 30 40 47DAYS FROM START OF DAYS FROM START OFNUTRIENT ADOmONS NUTRIENT ADDITIONS13.3121‘-. loj& I•Ce I3.- . —0—-a—0-— .3___SSo:.-uJc.10 17 23 30 40 3DAYS FROM START OFNUTRIENT ADDITIONSFigure 13. Numbers of adult insects by target phosphorus concentrationcoUected in emergence traps. All insects were captured betweenMay 19 and July 4. In each case, the mean number of adult insectsnumbers is plotted per date. Sample size (n) = 3 except at50 ug Pit where n = 1.(13.1) Chironomidae. (13.2) Trichoptera. (13.3) Plecoptera.468.2.2 Benthic insectsIncreasing phosphorus concentration resulted in an increase in total benthic insects countssampled at the end of the experiment. The relationship was neariy significant (y =349.85*log(P) + 1000.55, p = 0.06, r2 = 0.20, t05(2)11 = 2.05, power (1 - = 0.67, fortreatments 0 - 10 ug P1-1).Numbers of benthic Baetidae, Hydroptilidae, Nemouridae, and Perlodidae showed asignificant increase with increasing nutrient levels (Figure 14 and Table 6). Numbers of allfour of these insect taxa showed rapid increases at low phosphorus concentrations of 0 -2.5 g P. t’ and showed signs of saturation at 2.5 - 10 g P. l’. Other major insect taxacollected from the benthos showed no significant response to fertilization (Table 6).The number of taxa per sample (richness) of the trough benthos was not influenced bytreatment (p = 0.80, r2 = -0.08, t)j2 = 0.25). A comparison of the number of taxa persample in the control troughs to the natural stream showed no significant difference(p > 0.10, t05 = 1.52).The control trough benthos was comprised of four main families including Chironomidae,Philopotamidae, Baetidae, and Nemouridae which made up 82.6 - 99.1 percent of the totalinsects. While the benthos of shallow natural riffles near the mesocosm were comprised ofBaetidae, Chironomidae, Chloroperlidae, Elmidae and Leptophlebidae which made up 89.2percent of the total insects (Figure 15). Although the percent composition of the controltroughs and stream benthos appears different, the same taxa were at least present in bothtrough and stream benthos.474814.1 13.2400 300aa LuLuI-Lii30O_—.Go200 • OJ,Li. Lii 100OLuo 100zz0• 5 S 10 50 .52.5 5 10 50TARGET PH0SPH0P1.S CONCENTRATiON (uq P11)TARGET PHOSPHORUS CONCENTRATION (u P1)14.414.34050 ECaaLiiI—uJC,Lii 40 30hii.D 30 ...2OCU -‘0’:. .ZLLJ10 -I 101CC,z.5 2.5 s 10 50.0 .5 2.5 5 10TARGET PHOSPHCRS CONCENTRATION (ug P/I)TARGET PHOSPHORUS cONcENTRATIcH : P11)Figure 14. Number of insects collected from the benthos at the end of theexperiment on July 8 as a function of target phosphorusconcentration (ug Pt’). Linear regression lines of the data werefitted by least squares. See Table 7 for the results of regressionanalyses carried out on 0 - 10 ug P11 treatments for the same data.(14.1) Baeticlae. y = 152.06 + 89.96 * log(P).r2=0.443. p=.0O6.(14.2) Nemouridae. y=1l7.92 + 93.02 * log(P). =0.421. p0.0O8(14.3) Perlodidae. y=94l + 13.31 * log(P). r=0.402. pO.0l(14.4) Hydroptilidae. y=l3.95 + 10.26 * 10gW). rz0.16. p=0.0849Linear regression analyses relating the numbers (y) ofvarious insect taxa collected from the trough benthos collected atthe end of the experiment to target phosphorus concentration.Sample size (n) = 14 in all cases. Ordered with respect toincreasing p value.Insect Regression(2)TrophicFamily quationW t0•5(2),12 Relationship(Order)Perlodidae y = 16.29 - log(P) + 16.29 0.02 2.77 0.34 predators(Plecoptera)Nemouridae y = 108.54 * log(P) + 112.25 0.02 2.72 0.33 shredders,(Plecoptera) collectorsHydroptilidae y = 16.91 * log(P) + 14.01 0.02 2.58 0.30 scrapers,(Tricoptera) piercers,collectorsBaetidae y = 93.25 log(P) + 150.86 0.03 2.49 0.29 scrapers,(Ephemeroptera) collectorsPolycentropodidae y = 3.56 *log(P) + 5.40 0.07 1.96 0.18 collectors,(Tricoptera) some predatorsPhilopotamidae y = 70.45 * log(P) + 206.72 0.09 1.84 0.15 collectors(Tricoptera)Miscellaneous y = 0.21 * P + 4.68 0.15 1.54 0.09 scrapers,Ephemeroptera collectorsElmidae y = 0.25 * p + 5.35 0.26 1.18 0.03 collectors(ColeopteraChloroperlidae y = -0.26 * P + 4.82 0.31 -1.06 0.01 predators(Plecoptera)Hydropsychidae y = -0.47 * P + 14.01 0.39 -0.88 -0.02 collectors,(Tricoptera) some predatorsChironomidae y = 4.17 * P + 453.82 0.56 0.60 -0.05 various(Diptera)Miscellaneous y = 0.32 * P + 24.03 0.79 0.27 -0.08 variousDiptera“i Regression analyses included phosphorus treatments of 0-10 g Pl only.None of the regression analyses tested were significant at a “table-wide”a level accordingto the sequential Bonferroni technique (Rice 1988).Table 6.50__1000z075C1)(1)LU 50Ocnow00z0>-