Southwestern Association of Naturalistspossible that other benthic fishes such as small- mouth buffalo and river carpsucker perform similar roles. MATERIALS AND METHODS-Study Area-Lake

Southwestern Association of Naturalists

Feeding Ecology of Three Omnivorous Fishes in Lake Texoma (Oklahoma-Texas)Author(s): Keith B. GidoSource: The Southwestern Naturalist, Vol. 46, No. 1 (Mar., 2001), pp. 23-33Published by: Southwestern Association of NaturalistsStable URL: http://www.jstor.org/stable/3672370Accessed: 16/12/2010 14:00

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained athttp://www.jstor.org/action/showPublisher?publisherCode=swan.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

Southwestern Association of Naturalists is collaborating with JSTOR to digitize, preserve and extend access toThe Southwestern Naturalist.

http://www.jstor.org

http://www.jstor.org/action/showPublisher?publisherCode=swan

http://www.jstor.org/stable/3672370?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/action/showPublisher?publisherCode=swan

THE SOUTHWESTERN NATURALIST 46(1) :23-33 MARCH 2001

FEEDING ECOLOGY OF THREE OMNIVOROUS FISHES IN LAKE TEXOMA (OKLAHOMA-TEXAS)

KEITH B. GIDO*

University of Oklahoma, Biological Station and Department of Zoology, Norman, OK 73019

*Correspondent: [email protected]

ABSTRACT-Feeding ecology of 3 omnivorous fishes in a large southern United States reservoir was investigated to develop hypotheses on the potential functional roles of these species in this

ecosystem. I examined distribution, abundance, and diet of smallmouth buffalo (Ictiobus bubalus), river carpsucker (Carpiodes carpio), and gizzard shad (Dorosoma cepedianum) relative to the avail-

ability of potential resources during summer 1997 and 1998. In July and August, abundance of smallmouth buffalo was significantly greater at stations with depths of 10 m than at 1 m or 3 m stations. There was no significant difference in abundance of gizzard shad or river carpsucker by depth or month. Relative proportions of detritus and zooplankton in the diet varied among species. Smallmouth buffalo primarily ate copepods, whereas gizzard shad primarily ate detritus. The diet of river carpsucker was intermediate in relative proportions of zooplankton and detritus to that of smallmouth buffalo and gizzard shad. Of the 3 species examined, only gizzard shad showed a significant decline in dietary crude protein, phosphorus, and organic content over the summer. This corresponded to a decline in condition of gizzard shad. Overall, benthic invertebrates had a

heterogeneous distribution within the reservoir, and organic content of sediments was not different across sample stations. Relative importance to ecosystem functioning of these species, all of which are highly abundant in southern reservoirs, likely depends on species-specific feeding ecol-

ogy and environmental conditions.

RESUMEN-La ecologia de tres peces omnivoras en una presa del sur de Estados Unidos fue

investigada para desarrollar hip6tesis del papel funcional potencial en este ecosistema. Yo examine la distribuci6n, abundancia, y dieta de Ictiobus bubalus, Carpiodes carpio, y Dorosoma cepedianum, relativo a la abundancia potencial de los recursos durante los veranos de 1997 y 1998. En julio y agosto la abundancia de Ictiobus bubalus fue significativamente mayor en los sitios con profundi- dades de 10 metros que en los sitios con im o 3 m. de profundidad. No hay diferencias signifi- cativas en la abundancia de D. cepedianum o C. carpio por profundidad o por mes. Las proporciones relativas de detritus o zooplancton en la dieta fue variada entre las especies, Ictiobus bubalus come

principalmente copepodes mientras que D. cepedianum come detritus. La dieta de C. carpio fue intermedia en proporciones relativas de zooplancton y detritus a la de I. bubalus y D. cepedianum. De las tres especies examinadas, solo D. cepedianum exhibi6 una reducci6n dietetica significativa de proteina cruda, f6sforo, y contenido organico durante el verano. Esto corresponde a un declive en la condici6n de D. cepedianum. En general, los invertebrados bent6nico tienen una distribuci6n

heterog6nea dentro la presa y el contenido organico del sedimento no fue diferente a traves de los sitios de muestreo. La importancia relativa de estas especies en la funcionalidad del ecosistema, todas altamente abundantes en el lado sur del reservorio, probablemente depende de la relaci6n

ecol6gica de la alimentacion especie-especies y las condiciones ambiental.

Fish assemblages in southern reservoirs typ- ker, 1996). The degree they affect ecosystem ically are dominated by large-bodied benthic processes, however, depends on their abun- omnivores (Mensinger, 1971; Robison and Bu- dance, modes of feeding, and abiotic condi-

chanan, 1988). Such fishes can have impor- tions (e.g., nutrient loading from the water- tant effects on ecosystem processes such as shed and sedimentation rates). To assess po- nutrient cycling and bioturbation of sedi- tential roles of large-bodied omnivores in res-

ments (Lamarra, 1975; Brabrand et al., 1990; ervoirs, a requisite first step is understanding Stein et al., 1995; Drenner et al., 1996; Flec- where and when they forage and how this re-

THE SOUTHWESTERN NATURALIST 46(1):23-33 MARCH 2001

The Southwestern Naturalist

lates to resource availability (e.g., Power, 1997).

I examined feeding ecology of 3 omnivorous fish species (gizzard shad, Dorosoma cepedianum; smallmouth buffalo, Ictiobus bubalus, and river

carpsucker, Carpiodes carpio) in the context of the distribution and abundance of their food resources in Lake Texoma (Oklahoma-Texas). Omnivorous fishes, such as these, will switch to detritus if higher nutritional quality prey declines (Ahlgren, 1990a; Lobon-Cervia and Rin- con, 1994; Valladolid and Przybylski, 1996; Yako et al., 1996) and will grow best when detritus is supplemented with invertebrate prey (Mundahl and Wissing, 1987; Ahlgren, 1990b, Bowen et al., 1995). Thus, abundance and condition of these fishes may depend on spatial and temporal variation in benthic invertebrate abundance. In flood control reservoirs, benthic invertebrates may vary with wave exposure (Cooper, 1977), dissolved oxygen (Sublette, 1957; Cooper, 1980; Cooper and Knight, 1985), and particulate organic matter

(Vaughn, 1982). Moreover, detritus quality and

quantity may vary throughout lentic systems. For example, detritus from wind-exposed shoreline habitats maintains fathead minnows

(Pimephales promelas) better than detritus from

profundal habitats (Lemke and Bowen, 1998). The goal of this study was to correlate dis-

tribution, abundance, and diet of 3 omnivorous fishes to spatial and temporal variation in

quality and quantity of resources in Lake Te- xoma. This information was then used to de-

velop hypotheses regarding interspecific differences in potential effects of these fishes on material processing in reservoirs. Although studies previously have suggested gizzard shad can

play an important functional role in reservoir

ecosystems by processing detritus (Stein et al., 1995; Vanni, 1996; Schaus et al., 1997), it is

possible that other benthic fishes such as smallmouth buffalo and river carpsucker perform similar roles.

MATERIALS AND METHODS-Study Area-Lake Te- xoma is a 36,000 ha impoundment of the Washita and Red rivers on the Oklahoma-Texas border. Res- ervoir releases and resulting fluctuations in water level are primarily for hydropower and flood control. Near my study sites, Secchi depth transparency typically ranges from 100 to 125 cm, but can decrease to 15 cm during turbid inflow episodes (Mat- thews, 1984). Study sites were located ca. 35 km up-

lake from Denison Dam, within the Red River arm of Lake Texoma near the University of Oklahoma

Biological Station (Fig. 1). Nine sampling stations were established in 3 coves (3 stations per cove) located on the north shore of the reservoir. Within each cove, stations were located at depths of 1 m, 3 m, and 10 m. One and 3 m stations were within a

given cove and the 10 m stations were located di-

rectly outside the cove in the main body of the reservoir. All sites were located off-shore by a minimum of 20 m. Because depth profile, size, and wind ex-

posure were similar among the 3 coves, each was used as a replicate for statistical analyses.

Field Collections and Dietary Analysis-Adult fishes of the target species were collected monthly from June to August 1997 and 1998 at each station using ex-

perimental gill nets (46 m by 1.8 m) with mesh sizes

ranging from 51 to 101 mm bar-measure mesh. Gill nets were set with the lead line on the bottom dur-

ing daylight for 2 to 4 h. At each station (n = 9) a maximum of 2 individuals of each species (gizzard shad, river carpsucker, and smallmouth buffalo) was sacrificed for gut content analysis (i.e., maximum of 18 individuals of each species per sample date); re-

maining fish were counted and released. Intestines from sacrificed fish were immediately removed and

transported on ice to a freezer. For analysis of diet, gut contents were taken from

the esophagus (gizzard shad) or anterior quarter of the intestine (smallmouth buffalo and river carpsucker) and preserved in 70% ethanol. Esophageal contents of smallmouth buffalo and river carpsucker were not used because food items were rarely found in this region. Moreover, food items in the anterior

gut were intact and did not appear to have under-

gone much digestion. Gut contents from each fish were stirred to create a homogeneous mixture of food items; then a 1 ml subsample was taken from this mixture for dietary analysis. A preliminary in-

vestigation showed that replicate samples of this ho-

mogenate were not necessary because of little difference in percent volume of major food items (+5%) among replicates. The subsample was placed in a Sedgwick-Rafter counting slide and analyzed at 40X magnification. Relative volumes of food items for each sample were estimated by approximating the area occupied by individual items in 20 fields of view. Invertebrates were identified into major taxonomic groups as follows: phylum Rotifera; orders

Copepoda and Ostracoda; family Chironomidae; and genera Daphnia and Bosmina. Because all species had a relatively fine-grained diet, determination of the nature of vegetative food items was difficult. Therefore, I classified as detritus vegetative debris, algae, and amorphous organic matter.

Contents of the anterior one-fourth of the intestine (for all species) also were examined for crude

protein, total phosphorus, and percent organic mat-

24 vol. 46, no. 1

Gido-Feeding ecology of fishes in Lake Texoma

Washita River

Enlargement of study area

\ I

I I 10 km

N N I'f

Lake Texoma

FIG. 1-Location of the 9 sample stations in the Red River arm of Lake Texoma. Samples were taken at

depths of 1, 3, and 10 m in each of 3 coves.

ter to assess nutritional value of the diet. A maximum of five individuals of each species was examined each month; only fish with intestines more than 75% full were considered in these analyses. Intesti- nal contents were oven-dried at 60?C for 24 h and cooled in a desiccator. They were then ground with a mortar and pestle to homogenize the contents. To- tal Kjedahl nitrogen and total phosphorus were determined from a 0.25 g subsample of the ground contents that was digested in concentrated sulfuric acid at 440?C for 5 min. Total nitrogen was determined by the Nessler method and total phosphorus by the ascorbic acid method (American Public Health Association [APHA], 1985). Total nitrogen was converted to crude protein by a standard con- version factor (total nitrogen X 6.25). Organic content was estimated in the remainder of the sample by ash-free dry weight (AFDW) based on the difference in weight of dried samples from those com- busted in a muffle furnace at 550?C for 1 h. Al-

though crude protein and percent organic matter

may poorly reflect detritus nutritional quality because of the importance of non-protein amino acids

(e.g., Bowen, 1980), these measures should show dif-

ferences in diet due to relative proportions of invertebrate prey.

Benthic Invertebrates and Detritus-Abundance of benthic invertebrates and the organic fraction in benthic sediments were estimated from core sam-

ples. Core samples were taken with a combination of an acrylic tube (8 cm diameter) and an Eckman

dredge (15 by 15 cm). First, a sample of the benthic substratum was brought to the surface with the

dredge and a subsample was then taken from within the dredge with the corer before opening the bottom of the dredge. This allowed retrieval of core

samples from deep stations with only minimal dis- turbance of organisms and organic matter at the surface-water interface. Only the top 1 cm of each core

sample was retained for analysis. Two samples were taken at each station for each

sample date (n = 18) and stored at 4?C. In the lab-

oratory, a 5 g (wet-weight) subsample from each core was dried at 60?C for 24 h, and organic content was determined by combustion as previously described. The remaining sample was passed through a 210 Jm sieve to retain organic debris and macroinvertebrates. Macroinvertebrates were identified and

March 2001 25

' I


enumerated under a stereoscope at 20X magnification and placed into major taxonomic categories as previously described; with the exception of cladoc- erans, that were combined into one group. Herein, all invertebrates, including those associated with the sediment-water interface (i.e., Copepoda and Ostra- coda) are considered benthic invertebrates.

Sedimentation Rates-Because sedimentation of organic and inorganic materials can influence benthic resources (e.g., invertebrate abundance and deposition of phytoplankton and detritus), cylindrical sediment traps were placed on the bottom at each station concurrent with the gill net sampling in 1997 to estimate settling rates and organic fraction of sediments. Traps were retrieved ca. 48 h after deploy- ment. Each trap had a diameter of 9.9 cm and depth of 30.5 cm (height:diameter ratio > 3:1; Blomqvist and Kofoed, 1981) and was placed so the opening was 40 cm above the substratum. Before removal, each trap was capped in place (using SCUBA) and then brought to the surface. Sediment samples were allowed to settle for 1 h and then excess water was decanted. Dry weight of each sample was determined after drying at 60?C for 24 h, and percent organic content was determined by combustion (as described previously).

Data Analysis-Differences in abundance among depths, months, and years for each species of fish, major benthic invertebrate taxa, percent organic content in core samples, and sedimentation rates were tested using a repeated measures ANOVA with month as the repeated factor. Mean values from

paired samples were used for benthic invertebrates and percent organic content from core samples. Thus, for all variables there were 3 replicates (1 per cove) at each depth. Due to this low replication, statistical power for comparisons was low and subtle differences among treatment effects were probably not detected. Log (x + 1) transformations of fish and benthic invertebrate abundances were performed to maximize homogeneity of variances. Mul-

tiple comparisons among depths and months were analyzed with Tukey HSD tests (SPSS, 1996).

Principal components analysis (PCA) was used to characterize differences in diet among species and across sample dates (Crow, 1978). Prior to analysis, percent volume of major resource categories for each species was arcsine square root transformed to reduce deviation from normal distributions. Eigen- values and loadings were calculated using PC-ORD (McCune and Mefford, 1995) based on a correlation matrix of variables. A one-way ANOVA was used to detect differences among species in crude protein, total phosphorus, N:P ratio, and organic content of the diet. In addition, a general factorial ANOVA was used to test for differences in these variables between years and among months for each species.

A Fulton-type condition index (weight/length3;

Gizzard shad Im - 3m

T _ 10m 3

2-

1-

Ce iU 0

0 cc

-cr

0

.0 E z

U *. .- , I I

v II'I

Smallmouth buffalo 4-

3-

2 -

River carpsucker

2-

, 11n

0 ~ June July Aug June July

1997 1998

Aug

FIG. 2-Mean abundance (individuals per hour gill netting) of fish species taken from 3 coves in Lake Texoma by station depth, month, and year. Ver- tical bars represent 1 SE.

Anderson and Gutreuter, 1983) was used to assess condition of the three species throughout the summer. To increase sample size, lengths and weights of fish taken from a separate gill net survey in the same area of the reservoir (Gido, 1999) were used to cal- culate mean condition from June to August in 1996 and 1997. No analyses were performed in summer 1998 because of low sample sizes. Differences in condition among months were determined using a one- way ANOVA. Similarly, differences in sedimentation rates and percent organic contents of sediments among depths and months were examined using a general factorial ANOVA. All post hoc comparisons among depths and months were made by Tukey HSD tests (SPSS, 1996). All ANOVAs were performed using SPSS (1996).

RESULTS-Fish Abundance and Diet-Differ- ences in fish abundance were found across

depths and years, but not among months (Fig. 2). Mean abundance of gizzard shad was significantly higher in 1997 than in 1998 (F1,9 =

7.22, P = 0.025), but abundance did not vary with station depth. There was a significant difference in mean abundance among depths for smallmouth buffalo (F2,9 = 7.07, P = 0.014), but not between years. Although there was no difference in overall abundance among months for smallmouth buffalo, they were more evenly dispersed across depths in June

4

26 vol. 46, no. I


n

l c .E. . . 3 , 2 2

All species * June River -e n ,2 - C O

D G a July carpsucker

1 - 1 - A August

co

r - s

sJRr, ,

J-

3 -2 - 1 0 1 2 3 -3 -2 -1 0 1 2 3

O 3 . . 3 Q- Gizzard shad Smallmouth

2- 2 a\ / \ buffalo

-1 - - 1 -2

.-2

r-

C -3 ' -3

-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3

Copepoda Daphnia

< PC 1 (21.3%) > Detritus

FIG. 3-Differences in diet of 3 omnivorous fish species in Lake Texoma as revealed by the first 2 axes of a principal component analysis. Top left graph represents differences in diet among species: S = smallmouth buffalo, G = gizzard shad, R = river carpsucker. Other graphs represent monthly chang- es in diet for each of the 3 species.

but primarily occurred at 10 m depths in July and August. Because only 2 river carpsuckers were taken in 1997, comparisons of mean abundance among depths and months were made only for 1998 and were not significant (P > 0.10).

Principal components analysis revealed differences in diet among the three species,

though there was much overlap (Fig. 3). The first 3 axes accounted for 48% of the variation in diet among species (Table 1). Based on variable loadings, the first axis represented a contrast between individuals that consumed large volumes of cyclopoid copepods (hereafter re- ferred to as copepods) and Daphnia with those that consumed detritus. This axis identified both interspecific and seasonal differences in diet of these species. The second axis represented the relative proportions of ostracods, Bosmina, and terrestrial insects and primarily identified intraspecific variation in the diet of

gizzard shad. Axis 3 (not shown) represented a contrast between individuals that consumed

Daphnia and those that consumed zooplankton ephippia and Bosmina and also showed weak

interspecific differences between smallmouth buffalo and the other 2 species.

Averaged across months, copepods accounted for the greatest percent volume in the diet of smallmouth buffalo (x = 50.0%), whereas

gizzard shad consumed primarily detritus (x =

80.2%, Table 2). The diet of river carpsucker was intermediate between these species in percent volume of copepods (x = 32.3%) and detritus (x = 55.7%). In addition, all species tended to have a greater relative volume of co-

pepods in the diet in June (x = 40.5%) than in July or August (x = 26.2%).

A one-way ANOVA revealed significant differences among species in dietary crude pro-

TABLE 1-Eigenvalues and variable loadings for the first 3 principal component axes derived from an

analysis of relative volume of food items in the diet of 3 omnivorous fishes in Lake Texoma, 1997-1998. Asterisk indicates items considered important for that axis.

Axis

PC1 PC2 PC3

Eigenvalue 2.132 1.382 1.214 Percent variance explained 22.2 13.7 12.1 Food item

Copepoda -0.882* 0.153 -0.165

Daphnia -0.517* 0.065 0.454* Chironomidae -0.319 0.004 0.328 Ostracoda 0.105 -0.673* 0.082 Bosmina -0.206 -0.617* -0.466* Unknown zooplankton 0.161 -0.275 -0.315 Detritus 0.916* 0.112 0.216 Rotifer -0.297 0.096 0.211

Ephippia -0.131 0.313 -0.591* Terrestrial insect -0.222 -0.565* 0.326

March 2001 27


C~Co

cli ~c cnj Co

6 Co d;

C' C_

. -

0 C ?r06

0 0

0 6

Cl 101

C,I

6

00 cc

t> I O N

'

r- (0 P.^ CO

ccr 0

d ors

c,.

m'o-e-

* C

10

tr O c.

O

G

/- . .^ ^

Gld 0

_^r^SO s_ . ^_/ ?

r- 0s CO

Gur < f-

10

d o I

00 0C

I-- 0 0 0 in C3

d d o3j 0.3 cc 0.

2

z

6 m E

.-

Os

N

C,.

cc

t

06 cli

(X

CC -- 00 C CS q

Clr o00

0 6

_

co" co^ c\r in" ^ d; r-< 0 ^ d r- r-(

C-( Cf Cr1 C1 0 0 d; d ^ d d= d o~~~~o

o- o ^ t o o^

0C*(s

OI _r s o n

d' d d dj

73

- W 0 l C O Gl

? ? ?

C!UO~~~d., .

o f U 3 Q, H 5

* Smallmouth buffalo 30 - * Gizzard shad

z .A River carpsucker

20

10 _ ..

0 o

0.6 -

0.4

0.2-

0.0 18

15

12-

9

6

60*

40- ? "' "-, 20 -

O

Jun June Aug

1997

Jun June Aug

1998

FIG. 4-Differences in crude protein, total phosphorus, N:P ratio, and percent organic matter in the diet of 3 omnivorous fish species in Lake Texoma. Vertical bars represent 1 SE.

tein, total phosphorus, N:P ratio, and percent organic matter (P < 0.001; Fig. 4). Mean values for all variables were greatest in the diet of smallmouth buffalo, and least in gizzard shad. Excluding 1997 data because of low sample size, gut contents of river carpsucker had intermediate levels of all variables except N:P ratio. Low values of crude protein, phosphorus, and percent organic matter in the diet of gizzard shad were likely due to large volumes of sand rather than low protein content of detritus. No significant differences between years or among months for crude protein, total phosphorus, or percent organic matter were observed in the diet of smallmouth buffalo. How- ever, there was a significant year by month interaction for N:P ratio (F2,14 = 5.406, P = 0.018). This interaction occurred because of a decline in N:P into summer 1997 and an increase in N:P into summer 1998. Because of the low abundance of river carpsucker in 1997, only differences among months for 1998 were tested. No significant differences were found

28 vol. 46, no. I

QI z

a.

C

.

Cd

12

h0

Cd

"6

"6

ltt

Cd N

M 11

Ct3 "5 11

C

Ic-i

II

- 11

0 - _r

II

: 11

11

i;C

._

o

o Cz

06

ct

C

o

C

0.3

o

C

.,

O

o

o

56

o

o

CO

o

+1

u

"S

02

.s -S c^

(U 02 12

c

o ao

Q- (1)

0

Q. 5a


1.00

0.95 -

0.90 -

0.85 - 0

ur

x ^- 0.80

m 0.20- 0 L- m 0.18 -

0

m

u 0.1 -

c

, 0.12 - "0

o 0.36 -

0.32 -

Gizzard shad (n = 288)

a

Smallmouth buffalo (n = 142)

a

7 River carpsucker (n = 36)

a

a

a

E u,

10 .5 cp

6.0

0

E z

0.28- T -

0.24 -V//

0.20 June July August

FIG. 5-Mean condition of 3 omnivorous fish species in Lake Texoma across summer months for 1996 and 1997. Vertical bars represent 1 SE. Different let- ters above error bars represents significant difference between means.

for any of the variables. Dietary nutrient content of gizzard shad showed a significant effect of month for all measures of diet quality (P < 0.025). These differences were attributed to a general decline in all variables from June to August.

Mean condition of gizzard shad in 1996 and 1997 was significantly lower during July and August than in June (P < 0.05); however, no significant differences were observed for smallmouth buffalo or river carpsucker (Fig. 5).

Benthic Invertebrates and Detritus-Chirono- midae (41.2%), Copepoda (29.8%), and Ostra- coda (22.9%) accounted numerically for 94%

4000 Ostracoda 1 m

3000 ' 3m

2000 I

1000 ?

Chironomidae

6000 - I

4000 i

2000' li

Copepoda 8000

6000 | 4000 ?

2000 ?| o _&L ^ l - _ - June July

1997

Aug June July

1998

Aug

FIG. 6-Differences in abundance of 3 major taxa of benthic invertebrates in core samples taken from 3 coves in Lake Texoma by station depth, month, and year. Vertical bars represent 1 SE.

of the total individuals from core samples. All taxa showed heterogeneous distributions across station depths, months, or years (Fig. 6). Mean abundance of ostracods was not different (P > 0.05) among months, but varied significantly among depths (F2,10 = 7.509, P =

0.010) and years (F1,10 = 8.408, P = 0.016). Os- tracods were typically more abundant at 1 m and 3 m stations, and overall were more abundant in 1997 than in 1998. Mean chironomid abundance varied differently across months and years as shown by the significant interaction between these variables (F2,20 = 20.254, P < 0.001). Their abundance was greater inJune than in July or August and greater overall in 1997. Mean chironomid abundance in June was also higher at 1 m than at the 3 m or 10 m stations. Mean copepod abundance also varied significantly across months and depths (month by depth interaction, F4,20 = 9.030, P = 0.002). Copepods were most abundant at the 10 m stations, although this difference was only significant for July and August.

Organic Matter in Core Samples and Sedimenta- tion Rates-No significant differences in organic content of core samples were detected among months, depths, or years (P > 0.05). In general, mean percent organic matter in core samples was low (3.87%) and varied consider- ably among paired samples (CV = 115%). Ad-

z . . . . . . . . . . I

March 2001 29


700

- 600 - r 3m

E 500- ;

I 400- -

300- T i I

E 200-

Q)U) :: : (I 100- '' ' .

0

June July Aug

FIG. 7-Sedimentation rates taken from 3 coves on Lake Texoma across 3 depths in 1997. Vertical bars represent 1 SE.

ditionally, there were no differences (P < 0.05) in percent organic matter in sediment samples among months or depths, suggesting a relatively homogeneous distribution of organic matter in sediments. Sedimentation rates did vary significantly among depths (F2,13 = 8.389, P =

0.005) and months in 1997 (F2,13 = 7.837, P = 0.006; Fig. 7). Mean sedimentation rate was greatest in July, and, in all months, was significantly greater at the 3 m and 10 m stations than at 1 m stations. Thus, even though organic fraction of sediments appears homogeneous across habitats, the rates of deposition dif- fered.

DIscussION-Interspecific Differences in Feeding Ecology-In Lake Texoma, diets of gizzard shad, river carpsucker, and smallmouth buffalo dif- fered in invertebrate species composition, nu- trients, and organic content. These differences were partially due to the different feeding strategies of these fishes. Based on morphology and diet, gizzard shad can be classified as a scooper (Gerking, 1994), that scoops mud from the bottom and roughly sorts and dis- cards unwanted matter through the gill rakers. Smallmouth buffalo and river carpsucker are suction feeders and more selectively filter detritus and invertebrates through their gill rakers. Percent organic matter in gut contents of all fishes was higher than in core samples, in- dicating selective consumption of food items from the substrate. However, in comparison to the two catostomids, the percent organic content in the diet of gizzard shad was low (<20%), owing to consumption of large quan- tities of inorganic sand along with detritus. Ap-

parently, this species either compensates for the low nutritional quality of its food by in- creasing consumption rates or assimilation ef- ficiency (Grimm, 1988), or by supplementing its diet with small amounts of invertebrates when available (Yako et al., 1996).

Food items consumed by these species were similar to those reported previously (Walburg and Nelson, 1966; Tafanelli et al., 1971; Sum- merfelt et al., 1972; Pierce et al., 1981). How- ever, the percent volume of detritus in their diets was typically less. Other investigators have reported mean volumes of detritus of 68% in river carpsucker (Summerfelt et al., 1972) and greater than 65% in smallmouth buffalo (Ta- fanelli et al., 1971) in Oklahoma reservoirs. Corresponding values for the present study were 57% and 36%, respectively. Because the other studies were carried out over the entire year, it is possible that these differences are due to the greater proportion of zooplankton in the diet during summer. In any case, previ- ous studies agree with the findings of this one that both species are facultative detritivores and consume detritus when availability of invertebrates is low (i.e., July and August).

Differences in feeding strategies among fish species may influence their relative importance in reservoir ecosystems. In this study, gizzard shad ingested large amounts of detritus and inorganic sediments, whereas smallmouth buffalo and river carpsucker filtered invertebrates and detritus without ingesting much inorganic matter. Thus, gizzard shad would presumably process more materials and have a greater effect on fragmentation and decomposition of detritus, microorganisms, and algae in sediments in Lake Texoma. The effect of foraging by river carpsucker and smallmouth buffalo on fragmentation and decomposition of detritus is more likely through initial processing of detritus and sediments rather than direct consumption and passage of materials through their alimentary canal.

Spatial and Temporal Variation in Diet and Food Resources-Reservoir ecosystems can present a spatially and temporally heterogenous environ- ment for fishes. External inputs from the wa- tershed, wave action, and fluctuation in water level can influence deposition and oxidation of sediments and detritus. In addition, mid-summer succession of zooplankton (Threlkeld, 1986) and benthic invertebrate (Sublette,

30 vol. 46, no. 1


1957) assemblages creates seasonally available food resources for many reservoir fishes. In this study, I found that declines in invertebrate abundance, particularly copepods, at the 1 m and 3 m stations from June to August corre-

sponded to declines in relative volume of invertebrates in the diet of the 3 fish species. For

gizzard shad, organic content and crude protein declined from June to August and was fol- lowed by a decline in condition. Thus, at least for gizzard shad, food quality and quantity declines into the summer, and they shift to a diet with lower nutritional value (i.e., detritus).

Numerous studies have shown that omnivorous fishes will switch to a lower quality diet when invertebrate abundance declines (Bra- brand, 1985; Mundahl and Wissing, 1988; Ahl-

gren, 1990a; Lobon-Cervia and Rincon, 1994; Ahlgren, 1996). Mundahl and Wissing (1987) found a similar pattern in an Ohio reservoir, where gizzard shad switched from a mixed zoo-

plankton and detritus to a primarily detritus diet, resulting in lower growth and condition

during summer. Ahlgren (1990b) also reported that juvenile white sucker (Catostomus commersoni) fed only detritus lost weight, whereas those fed invertebrates gained weight. Bowen et al. (1995) suggested that, because of limited

availability of invertebrates, many fish consume detritus to supplement their diet; however, nutritional quality of this resource is not ade-

quate to sustain growth and reproduction. Whereas the nutrient and organic contents

of the diet, along with condition of gizzard shad, declined from June to August, there was no significant decline in these parameters for smallmouth buffalo or river carpsucker. Even

though the relative volume of invertebrates in the diet of these species appears to decrease into the summer, they still may be able to main- tain condition by supplementing their diet with detritus (Bowen et al., 1995). Moreover, smallmouth buffalo likely selected deeper habitats during July and August because of the

greater abundance of copepods in these habitats.

Differences in nutritional quality of detritus in sediments among habitats may also influence distribution and abundance of detritivorous fishes. Lemke and Bowen (1998) showed that nutritional quality of detritus in sediments was greater in areas that were exposed to tur- bulence from waves than in profundal zones

sheltered from waves. Bowen (1984) also showed that the condition of male tilapia (Sa- rotherodon) was lowest in habitats with low quality foods. Although I did not rigorously examine nutritional quality of detritus in sediments

(e.g., examine amino acids), I did show percent organic matter in core samples and fresh-

ly deposited sediments did not vary across

depths. This homogeneous distribution of organic matter in upper layers of sediments appears to be due to deposition of sediments of similar organic composition. Thus, even

though spatial differences in benthic invertebrates occurred, organic matter in sediments was relatively constant across habitats. Homo- geneous distribution of organic matter in sediments may correspond to broad spatial distribution of gizzard shad, and perhaps river carpsucker, that have higher proportions of detritus in their diet. More detailed examination of detritus quality (e.g., amino acids and hydro- lysis-resistant organic matter) would be neces-

sary to determine if detritus quality affected distribution of these species.

Because benthic fishes must process sediments to attain nutritionally important food material (Minckley et al., 1970; Mundahl, 1991), rate of sediment deposition presumably can influence their effect on benthic communities. For example, Mundahl (1991) showed that gizzard shad processed <4% of the sediments deposited in an Ohio reservoir. He con- cluded that, because of high rates of deposition of sediments, foraging by gizzard shad should have little effect on benthic communities. Typical of many reservoirs (Neel, 1966), sedimentation rates in Lake Texoma were

high. At the scale of the whole reservoir, Lake Texoma has lost >11% of its initial storage ca-

pacity due to sediment deposition between 1942 and 1985, with the greatest areas of de-

position in the riverine portions of the reservoir (H. Hartwell, United States Army Corps of

Engineers, pers. comm.). Sediment trap data from this study further suggests sediment de-

position also increases with depth at the scale of individual coves. Therefore, these spatial differences in sedimentation will likely influence the impact these species have on benthic invertebrate assemblages.

Conclusion-The success of omnivorous species such as gizzard shad, smallmouth buffalo, and river carpsucker in many southern reser-

March 2001 31


voirs is likely due to their ability to switch to lower quality food items when invertebrate

prey are in low abundance (e.g., Cherry and

Guthrie, 1975) and the ability of adults to evade predation because of their large body size (Stein et al., 1995). Interspecific differences in distribution, abundance, and foraging be- havior of these species along with variable abiotic conditions likely influence their relative

importance in reservoir ecosystems. This study suggests gizzard shad should have the largest per capita effect on ecosystem processes given: 1) low organic content in their diet and presumably high rates of sediment processing; and

2) consumption and processing of detritus that would otherwise be locked in sediments.

Assistance in the field and laboratory was provided by R. Durtche, S. Gido, W. Luttershmitt, W. Mat- thews, J. Schaefer, and W. Wolfinbarger. I also thank R. Durtche, E. Marsh-Matthews, B. Narin, and R.

Smiley for thoughtful discussions. Earlier versions of this manuscript greatly benefitted from comments

by L. Canter, M. Kaspari,J. Schaefer, W. Shelton, W. Matthews, and C. Vaughn. Partial funding for this

project was provided from the Graduate Student Senate of the University of Oklahoma. This research was in partial fulfilment of a Ph.D. at the University of Oklahoma.

LITERATURE CITED

AHLGREN, M. O. 1990a. Diet selection and the con- tribution of detritus to the diet of the juvenile white sucker (Catostomus commersoni). Canadian Journal of Fisheries and Aquatic Sciences 47:41- 48.

AHLGREN, M. O. 1990b. Nutritional significance of facultative detritivory to the juvenile white sucker (Catostomus commersoni). Canadian Journal of Fisheries and Aquatic Sciences 47:49-54.

AHLGREN, M. 0. 1996. Selective ingestion of detritus

by a north temperate omnivorous fish, the juvenile white sucker, Catostomus commersoni. Environ- mental Biology of Fishes 46:375-381.

ANDERSON, R. O., AND S.J. GUTREUTER. 1983. Length, weight, and associated structural indices. In: Niel- sen, L. A., and D. L. Johnson, editors. Fisheries

techniques. American Fisheries Society, Bethes- da, Maryland. Pp. 283-300.

APHA (AMERICAN PUBLIC HEALTH ASSOCIATION). 1985. Standard methods for the examination of water and wastewater, Sixteenth ed. American Public Health Association, Washington, D.C.

BLOMQVIST, S., AND C. KOFOED. 1981. A review of sediment traps in aquatic environments. Limnology and Oceanography 26:585-590.

BOWEN, S. H. 1980. Detrital nonprotein amino acids are the key to rapid growth of tilapia in Lake Valencia, Venezuela. Science 207:1216-1218.

BOWEN, S. H. 1984. Differential habitat utilization by sexes of Sarotherodon mossambicus in Lake Valen- cia, Venezuela: significance for fitness. Journal of Fish Biology 24:115-121.

BOWEN, S. H., E. V. LUTZ, AND M. O. AHLGREN. 1995.

Dietary protein and energy as determinants of food quality: trophic strategies compared. Ecol-

ogy 76:899-907. BRABRAND, A. 1985. Food of roach (Rutilus rutilus)

and ide (Leusiscus idus): significance of diet shift for interspecific competition in omnivorous fishes. Oecologia 66:461-467.

BRABRAND, A., B. A. FAAFENG, AND J. P. M. NILSSEN. 1990. Relative importance of phosphorus supply to phytoplankton production: fish excretion ver- sus external loading. Canadian Journal of Fish- eries and Aquatic Sciences 47:364-372.

CHERRY, D. S., AND R. K. GUTHRIE. 1975. Significance of detritus or detritus-associated invertebrates to fish production in a new impoundment. Journal of the Fisheries Research Board of Canada 32: 1799-1804.

COOPER, C. M. 1977. Abundance and production of littoral and profundal benthic fauna in a flood control reservoir. Proceedings of the Mississippi Chapter of the American Fisheries Society 1:25- 33.

COOPER, C. M. 1980. Effects of abnormal thermal stratification on a reservoir benthic macroinver- tebrate community. American Midland Naturalist 103:149-154.

COOPER, C. M., AND L. A. KNIGHT, JR. 1985. Macro- benthos-sediment relationships in Ross Barnett Reservoir, Mississippi. Hydrobiologia 126:193- 197.

CROW, M. E. 1978. Multivariate statistical analysis of stomach contents. In: Lipovsky, S. J., and C. A. Simenstad, editors. Fish food habits studies proceedings of the second Pacific northwest techni- cal workshop. Washington Sea Grant Publication, Seattle. Pp. 87-96.

DRENNER, R. W., J. D. SMITH, AND S. T. THRELKELD. 1996. Lake trophic state and the limnological effects of omnivorous fish. Hydrobiologia 319:213- 233.

FLECKER, A. S. 1996. Ecosystem engineering by a dominant detritivore in a diverse tropical stream.

Ecology 77:1845-1854. GERKING, S. D. 1994. Feeding ecology of fish. Aca-

demic Press, San Diego, California. GIDO, K. B. 1999. Ecosystem effects of omnivorous

fishes in Lake Texoma (Oklahoma-Texas). Un- published Ph.D. dissertation, University of Oklahoma, Norman.

GRIMM, N. B. 1988. Feeding dynamics, nitrogen bud-

32 vol. 46, no. 1


gets, and ecosystem role of a desert stream om-

nivore, Agosia chrysogaster (Pisces: Cyprinidae). Environmental Biology of Fishes 21:143-152.

LAMARRA, V. A., JR. 1975. Digestive activities of carp as a major contributor to the nutrient loading of lakes. Verhandlungen der Internationalen Verei-

nigung ffir Limnologie 19:2461-2468. LEMKE, M.J., AND S. H. BOWEN. 1998. The nutritional

value of organic detrital aggregate in the diet of fathead minnows. Freshwater Biology 39:447- 453.

LOBON-CERVIA, J., AND P. A. RINCON. 1994. Trophic ecology of red roach (Rutilus arcasii) in a seasonal stream: an example of detritivory as a feeding tactic. Freshwater Biology 32:123-132.

MATTHEWS, W. J. 1984. Influence of turbid inflows on vertical distribution of larval shad and freshwater drum. Transactions of the American Fish- eries Society 113:192-198.

MCCUNE, B., AND M.J. MEFFORD. 1995. PC-ORD: Mul- tivariate analysis of ecological data, Version 2.0.

MjM Software Design, Gleneden Beach, Oregon. MENSINGER, G. C. 1971. Oklahoma commercial fish-

eries summary 1961-1969. Proceedings of the Oklahoma Academy of Sciences 51:23-28.

MINCKLEY, W. L., J. E. JOHNSON, J. N. RINNE, AND S.

E. WILLOUGHBY. 1970. Foods of buffalofishes, ge- nus Ictiobus, in central Arizona reservoirs. Trans- actions of the American Fisheries Society 99:333- 342.

MUNDAHL, N. D. 1991. Sediment processing by gizzard shad, Dorosoma cepedianum (Lesueur), in Ac- ton Lake, Ohio, U.S.A. Journal of Fish Biology 38:565-572.

MUNDAHL, N. D., AND T. E. WISSING. 1987. Nutritional

importance of detritivory in the growth and condition of gizzard shad in an Ohio reservoir. En- vironmental Biology of Fishes 20:129-142.

MUNDAHL, N. D., AND T. E. WISSING. 1988. Selection and digestive efficiencies of gizzard shad feeding on natural detritus and two laboratory diets. Transactions of the American Fisheries Society 117:480-487.

NEEL, J. K. 1966. Impact of reservoirs. In: Frey, D.

G., editor. Limnology in North America. Univer-

sity of Wisconsin Press, Madison. Pp. 575-593. PIERCE, R.J., T. E. WISSING, AND B. A. MEGREY. 1981.

Aspects of the feeding ecology of gizzard shad in Acton Lake, Ohio. Transactions of the American Fisheries Society 110:391-395.

POWER, M. E. 1997. Estimating impacts of a dominant detritivore in a neotropical stream. Trends in Ecology and Evolution 12:47-49.

ROBISON, H. W., AND T. M. BUCHANAN. 1988. Fishes of Arkansas. University of Arkansas Press, Fayette- ville.

SCHAUS, M. H., M.J. VANNI, T. E. WISSING, M. T. BRE-

MIGAN, J. E. GARVEY, AND R. A. STEIN. 1997. Nitro-

gen and phosphorus excretion by detritivorous

gizzard shad in a reservoir ecosystem. Limnology and Oceanography 42:1386-1397.

STATISTICAL PACKAGE FOR THE SOCIAL SCIENCES. 1996. SPSS Base 7.0 for Windows. SPSS Inc., Chicago, Illinois.

STEIN, R. A., D. R. DEVRIES, AND J. M. DETTMERS. 1995. Food-web regulation by a planktivore: ex-

ploring the generality of the trophic cascade hy- pothesis. Canadian Journal of Fisheries and

Aquatic Sciences 52:2518-2526.

SUBLETTE,J. E. 1957. The ecology of the macroscopic bottom fauna in Lake Texoma (Denison Reser-

voir), Oklahoma and Texas. American Midland Naturalist 57:371-402.

SUMMERFELT, R. C., P. E. MAUCK, AND G. MENSINGER.

1972. Food habits of river carpsucker and freshwater drum in four Oklahoma reservoirs. Pro-

ceedings of the Oklahoma Academy of Sciences 52:19-26.

TAFANELLI, R., P. E. MAUCK, AND G. MENSINGER. 1971. Food habits of bigmouth and smallmouth buffalo from four Oklahoma reservoirs. Proceedings of the Annual Conference of Southeastern Associa- tion of Fish and Wildlife Agencies 24:649-658.

THRELKELD, S. T. 1986. Resource-mediated demo-

graphic variation during the midsummer succession of a cladoceran community. Freshwater Bi-

ology 16:673-683. VALLADOLID, M., AND M. PRZYBYLSKI. 1996. Feeding

relations among cyprinids in the Lozoya River

(Madrid, central Spain). Polskie Archiwum

Hydrobiologii 43:213-223. VANNI, M. J. 1996. Nutrient transport and recycling

by consumers in lake food webs: implications for

algal communities. In: Polis, G. A., and K. O. Winemiller, editors. Food webs: integration of

patterns and dynamics. Chapman and Hall, New York. Pp. 81-95.

VAUGHN, C. C. 1982. Distribution of chironomids in the littoral zone of Lake Texoma, Oklahoma and Texas. Hydrobiologia 89:177-188.

WALBURG, C. H., AND W. R. NELSON. 1966. Carp, river

carpsucker, smallmouth buffalo and bigmouth buffalo in Lewis and Clark Lake, Missouri River. Bureau of Sport Fisheries and Wildlife, Research

Report 69, Washington, D.C. YAKO, L. A., J. M. DETTMERS, AND R. A. STEIN. 1996.

Feeding preferences of omnivorous gizzard shad as influenced by fish size and zooplankton den-

sity. Transactions of the American Fisheries So-

ciety 125:753-759.

Submitted 7July 1999. Accepted 29 December 1999. Associate Editor was David R. Edds.

33 March 2001

Documents

Southwestern Association of Naturalistspossible that other benthic fishes such as small- mouth buffalo and river carpsucker perform similar roles. MATERIALS AND METHODS-Study Area-Lake