|
342
<br />PAPOULIAS AND MINCKLEY
<br />Strauss's (1979, 1982) linear index (Li) of selec-
<br />tivity was used to evaluate selection for kinds and
<br />size-classes of foods. This index, L; = ri - p;,
<br />compares the relative abundance of prey i in the
<br />gut (r;) of a predator with the relative abundance
<br />of that prey in the environment (pi).
<br />Ponds were drained and fish were recovered 18-
<br />19 April 1985. Percentage survival was calculated
<br />for each pond based on initial stocking. Larval
<br />tiger salamanders Ambystoma tigrinum were
<br />abundant in some ponds and were also recovered
<br />because they are potential competitors of fish lar-
<br />vae. The number and biomass of salamanders was
<br />calculated for each pond.
<br />Limnological data, fish dietary data, survival,
<br />fish length and weight data, and salamander bio-
<br />mass data were analyzed by analysis of variance
<br />and Tukey's HSD (honestly significant difference)
<br />procedure. Strauss's linear index was tested for
<br />among-treatment differences by regression anal-
<br />ysis and comparisons of slopes, and for differences
<br />from zero by Student's t-test. Loss of samples ne-
<br />cessitated combination of data from weeks 2 and
<br />3 to create adequate sample sizes for some statis-
<br />tics. All statements of significant differences are at
<br />the 0.05 level. Means in text are ± 1 SE, where
<br />appropriate. All data manipulations and analyses
<br />were performed with Lotus, Systat, and Sigmaplot
<br />software.
<br />Results
<br />Limnological Data
<br />Mean surface water temperatures increased from
<br />12°C in February to 17°C in April. Temporal pat-
<br />terns of chlorophyll were similar in all treatments,
<br />peaking within the first 2 weeks after ponds were
<br />filled, first in high-level fertilizer treatments and
<br />later in those with less or no fertilization, then
<br />declining thereafter (Table 1). Mean chlorophyll
<br />concentrations were generally greater in highly fer-
<br />tilized than in unfertilized ponds; however, vari-
<br />ation was great and only during week 1 were dif-
<br />ferences among treatments significant.
<br />Numbers and biovolumes of invertebrates in
<br />high-fertilization treatments began to increase 1-
<br />3 weeks after ponds filled, peaking at week 6 (Ta-
<br />ble 1). Numbers and biovolumes for low-fertiliza-
<br />tion treatments were consistently low during the
<br />experimental period. Numbers of invertebrates
<br />peaked at week 4 and gradually declined for the
<br />remainder of the experiment in medium-treat-
<br />ment ponds, but biovolume remained low and
<br />similar to those in low-fertilization ponds. Com-
<br />pared with zooplankton concentrations in natural
<br />waters, all our pond densities were relatively low.
<br />Highly fertilized ponds supported the greatest
<br />numbers and biomasses, averaging 43.3 organ-
<br />isms/L and 2,455 mm3/m3. Medium- and low-
<br />fertilization treatments averaged 23.7 organ-
<br />isms/L and 354 mm'/m3, and 12.5/L and 257
<br />mm3/m3, respectively. Statistically, numbers in
<br />highly fertilized ponds were greater than those in
<br />low-fertilization treatments in weeks 2 and 3 com-
<br />bined and in week 4. High-and medium-fertiliza-
<br />tion treatments differed only for combined weeks
<br />2 and 3. Medium-fertilization treatments had sig-
<br />nificantly greater numbers of invertebrates than
<br />low-fertilization treatments in week 4. Biovol-
<br />umes did not differ significantly among treatments
<br />on any date.
<br />In highly fertilized ponds, cladocerans, uniden-
<br />tified invertebrate eggs, and copepod nauplii and
<br />adults dominated the zooplankton in decreasing
<br />order of numerical abundance (Figure 1). Rotifers,
<br />nauplii, and cladocerans dominated in medium-
<br />fertilization treatments, as did nauplii, ostracods,
<br />cladocerans, and rotifers in low-fertilization treat-
<br />ments. On average, numbers and volumes of cla-
<br />docerans, copepods, and eggs were statistically
<br />greater in high- than in medium- and low-fertil-
<br />ization treatments, but no other relationships were
<br />apparent.
<br />Invertebrate size-classes had similar distribu-
<br />tions in the three treatments, with small organ-
<br />isms predominating (Figure 2). However, animals
<br />larger than 0.2 mm wide became more abundant
<br />first in high- and later in low-fertilization ponds.
<br />There was greater diversity of invertebrate body
<br />widths in high- and medium-fertilization treat-
<br />ments than in low-fertilization treatments. Rela-
<br />tive distributions of biovolumes of individual or-
<br />ganisms in the various size-classes also diversified
<br />earliest in highly fertilized ponds, but tended to
<br />be similar in all treatments by week 4. Greatest
<br />diversity was nonetheless consistently in high-fer-
<br />tilization treatments. From weeks 3 through 6,
<br />organisms 0.4 mm or wider were more numerous
<br />and averaged greater in biovolume in high-fertil-
<br />ization treatments than in others.
<br />Fish Data
<br />Survivorship and growth. -After 8 weeks there
<br />were no statistical differences in survival of larval
<br />razorback suckers among low-, medium-, and high-
<br />fertilization treatments (67.4 ± 15.7%, 89.8 ±
<br />6.8%, and 77.0 ± 8.1%, respectively; F = 1. 60, P
<br />= 0.26; grand mean, 77.5 ± 6.3%). However, lar-
|