Introduction

In the 1970’s two invasive carp species were introduced to North America for use in aquaculture (Freeze and Henderson 1982). These two invasive carp (Silver Carp Hypophthalmicthys molitrix and Bighead Carp Hypophthalmicthys nobilis) will be referred to as Asian carp. Shortly after this introduction, a series of floods occurred allowing them to enter the Mississippi River system and expand their range (Kolar et al. 2007). Since this release they have expanded to the Missouri, Mississippi, Illinois and Ohio Rivers (Kolar et al. 2007) and are threatening to enter the Great Lakes via the Chicago Shipping canal (Conover et al. 2007). Once established in a new range, their populations increase exponentially (Chick and Pegg 2001; Irons et al. 2007; Valery et al. 2008; Sass et al. 2010; Hayer et al. 2014). Specifically, Silver Carp have reached dense populations, some with catches as high as nearly 50% of the total (Chick and Pegg 2001; Irons et al. 2007; Valery et al. 2008; Sass et al. 2010; Hayer et al. 2014). These dense populations can result in large reproductive events if spring flood conditions allow (Lohmeyer and Garvey 2009). These events often result in high densities of Young-of-Year YOY Silver Carp entering the system (Lohmeyer and Garvey 2009). While we understand these massive fish schools can devastate the plankton community (Schrank et al. 2003; Sampson 2005; Irons et al. 2007; Sampson et al. 2009; Wanner and Klumb 2009b; Hayer et al. 2014) little is known about their interactions with larger piscivorous fishes.

These massive schools of small Silver Carp seem to be an obvious prey source for the common predators of the Mississippi River system. Field observations suggest that native piscivores are consuming Silver Carp but the rate at which this consumption occurs is unknown.

Therefore, the goal of this research was to conduct a field study to assess the diets of native piscivores in the Mississippi River. Once selection or avoidance was identified, a controlled laboratory experiment was conducted to determine if common predators (White Bass Morone chrysops and Largemouth Bass Micropterus salmoides) would select or avoid this novel prey in the presence of two native prey fishes (Gizzard Shad Dorosoma cepedianum and Emerald Shiner Notropis atherinoides). Understanding how predator-prey interactions occur in a controlled laboratory experiment may provide insight to trends observed in the field. We hypothesize that Silver Carp are more effective than native prey fish at avoiding predation and therefore have a competitive advantage over the native prey fish in a controlled environment. If our hypothesis is correct then the ratio of first prey consumed will be significantly different than an expected ratio of 1:1:1.

Methods

Field study

The field study component of this project occurred in Pool 26 of the Mississippi River, just north of St. Louis, MO and the Open River Reach of the Mississippi River stretching from Grand Tower, IL to the confluence with the Ohio River near Cairo, IL. These two reaches are separated by a physical barrier, Lock and Dam 26, and differ in physical structure. Pool 26 is an impounded reach with lentic conditions while the Open River Reach is unimpounded and therefore consists of lotic conditions. Both reaches contain reproducing populations of Silver Carp and Bighead Carp collectively referred to as Asian carp.

Native piscivores were sampled in 2015 from June to October at the Open River Reach and Pool 26 of the Mississippi River. Standard Long Term Resource Monitoring (LTRM) protocols for daytime electrofishing were used to collect fish (Gutreuter et al. 1995). Each electrofishing run consisted of fifteen-minute duration covering a rectangular area of about two hundred by thirty meters. Following each run, predatory fish were measured to the nearest mm (TL) and weighed to the nearest gram. Gut contents were extracted via gastric lavage and immediately preserved in 95% Ethanol. In the lab, contents were inspected under a dissecting microscope and identified to lowest possible taxa.

Environmental abundances were estimated with LTRM catches of fishes less than 80 mm in the specific reach within one week of the sampling event using a complement of gears. LTRM gears included electrofishing, fyke nets (1.8 cm mesh, 15 by 1.3 m leads, 0.9 by 1.8 m frame and six 0.9 m diameter hoops), mini fyke nets (3 mm mesh, 4.5 by 0.6 m leads, 0.6 by 1.2 m frame and two 0.6 m diameter hoops), hoop nets (3 m long and 1.8 cm mesh with seven 0.5 m diameter hoops baited with 3 kg of soybean cake) and bottom trawls (4.8 m slingshot balloon; Gutreuter et al. 1995). These gears are used at stratified random and fixed–sites throughout each reach from June to October.

Using the relative proportions of prey fish in the diets and relative proportions of LTRM catches, Ivlev’s electivity selection index was calculated. Ivlev’s index is calculated as:

$$ {E}_i=\frac{\left({r}_i-{P}_i\right)}{\left({r}_i+{P}_i\right)}, $$

where r i is the abundance of the prey item in the diet, P i is the abundance of prey item in the environment and E i  = index ranging from −1 to 1 (Ivlev 1961). Mean index was calculated for each species and prey fish by reach.

Controlled laboratory experiments

Adult piscivorous fish were collected with daytime electrofishing from the Mississippi River including White Bass and Largemouth Bass then acclimated to lab conditions until fish would readily feed. Lab acclimation and trials were conducted in two rectangular water tanks (3 m long by 1 m wide with water filled to a depth of 0.5 m high divided by two barriers into three 1 m squares). Prey fish included Gizzard Shad, Emerald Shiner and Silver Carp were also collected from the Mississippi River by electrofishing. No prey fish exceeded eight cm in total length to ensure all fish could be consumed by the predators. Each trial consisted of three prey fish of each species (all within 10 mm of each other to prevent size selection) being placed with the predator. The order in which the predator consumed the prey fish was recorded. Prey selection (first prey item consumed) by each predator species was analyzed by using Chi-square Goodness-of-fit tests with an alpha level of 0.10.

$$ {x}^2=\sum \frac{{\left( observed- expected\right)}^2}{expected} $$

Results

Field study

Diets were collected from a total of three hundred and nine fish (Open River n = 190; Pool 26 n = 119) representing forty species. The majority of piscivores caught were White Bass (n = 50), Shortnose Gar (Lepisosteus platostomus; n = 50), Largemouth Bass (n = 38), Flathead Catfish (Pylodictis olivaris; n = 30), Channel Catfish (Ictalurus punctatus; n = 19) and Longnose Gar (Lepisosteus osseus; n = 8). Of the 309 gut contents assessed, 189 were empty. Of the remaining 120 that were not empty, 92 contained fish. Many non-fish prey items were observed in the diets including crawfish, aquatic invertebrates and bivalves. While these items appear to be important to these predators their concentrations or abundance in the environment is unknown so this study focused on the fish consumed.

In Pool 26 (Fig. 1a), Longnose Gar showed a selection for Asian carp (0.12) but a stronger selection for Gizzard Shad (0.28). Shortnose Gar (0.02) and White Bass (0.03) showed very little selection for Asian carp and Channel Catfish (−3.0E-7), Flathead Catfish (−4.1E-5) and Largemouth Bass (−4.8E-7) showed no selection for or against Asian carp. In the Open River (Fig. 1b), Asian carp were avoided by Largemouth Bass (−0.08), Longnose Gar (−0.24), Shortnose Gar (−0.50), and White Bass (−0.16). The strongest avoidance of Asian carp was by Shortnose Gar. Largemouth Bass actually showed a positive selection for Grass Carp (0.23) while Shortnose Gar selected for Gizzard Shad (0.15).

Fig. 1
figure 1

Mean Selection value (± SE) for the most common predators (CNCF = Channel catfish, FHCF = Flathead catfish, LMBS = Largemouth bass, LNGR = Longnose gar, SNGR = Shortnose gar and WTBS = Whitebass) on the common prey fish (Asian carp = both Silver and Bighead carp, Grass carp, Gizzard shad, Shiners species and Centrarchid species) in Pool 26 (A) and the Open River reach (B)

Full size image

Controlled laboratory experiments

Over the course of our evaluation, 57 trials were completed. Twenty-eight of those trials were completed using White Bass and twenty-nine completed using Largemouth Bass. The order in which prey fish were consumed by White Bass showed significant (χ2 = 6.93, df = 2 and p = 0.03) selection (Fig. 2). Silver Carp was consumed first in only 10.7% of White Bass trials (three out of twenty-eight trials as compared to the expected values 9.33), less than any other prey item. The order in which prey fish were consumed by Largemouth Bass showed (χ2 = 5.03, df = 2 and p = 0.08) selection (Fig. 2). Silver Carp was consumed first in only 13.8% of Largemouth Bass trials (four out of twenty-nine trials as compared to the expected value of 9.66), less than any other prey item.

Fig. 2
figure 2

Percent each prey fish (ERSN = Emerald shiner, GZSD = Gizzard shad and SVCP = Silver Carp) was consumed first for both White Bass and Largemouth Bass. The order in which prey fish were consumed by White Bass showed significant (χ2 = 6.93, df = 2 and p = 0.03) selection and the order in which prey fish were consumed by largemouth Bass showed (χ2 = 5.03, df = 2 and p = 0.08) selection

Full size image

Discussion

The frequency in which Silver Carp were consumed first differed from what we would expect without selection, one possible explaination for this is that Silver Carp may be superior at evading predation by White Bass and Largemouth Bass. If small Silver Carp are truly superior at evading predation, this could be an explanation as to why they have expanded so rapidly since their introduction. Why Silver Carp were consumed first less often than the other two prey species is not completely understood but may result from a different body shape or superior evasive movements. The use of an Ivlev’s index to show selection was shown cause issues with some non-normal datasets and sample size variation (Strauss 1979) and an alternative method can be used to alleviate this (Chappell & Smith 2016). However our results tend to show either no selection for or against Asian carp by predators and in one case a strong selection against.

It appears that the trend we established in the controlled experiment was supported by the field study. Not only were Asian carp selected against but a different carp (Grass Carp) was selected for by Largemouth Bass. It also appears that this selection against Asian carps can be expanded to other predatory species. Specifically, the Gar species appear to be consuming the most Asian carp but still selecting against them in Pool 26 (A) and the Open River reach (B).

Understanding that Asian carp are not selected for by the native predators may help explain why Silver Carp populations can expand at such exponential rates. It does appear that all these native predators do consume Asian carp under the right conditions and bolstering these native predator populations could aid in Asian carp control.

The interactions between predator and prey can be very complex and rely on many factors outlined by the optimal foraging theory (Charnov 1976; Pyke 1984). It appears that when factors like prey availability, predator size and prey size are controlled there still exists an avoidance of Silver Carp by Largemouth Bass and White Bass. This suggests that some behavioral differences may be playing a key role in prey selection. Silver Carp may be superior in burst speed and awareness of predator allowing them to escape predation more often than the Emerald Shiner and Gizzard Shad. Further research is required to better understand these relationships and how they can be used in the management of Asian carp.